
privacy-impact-assessment-skills
18 DPIA and PIA skills: GDPR Art. 35, risk scoring, stakeholder consultation, threshold screening, mitigation planning
- version
- 1.0.0
- GitHub stars
- 340
Install
/plugin install privacy-impact-assessment-skills@privacy-data-protection-skillsRun it inside Claude Code after adding the marketplace with /plugin marketplace add mukul975/Privacy-Data-Protection-Skills.
Contents
Discovered from sourceSkills 18
Conducting AI System Privacy Assessment
Overview
AI systems that process personal data require a combined privacy and conformity assessment addressing both GDPR obligations and the EU AI Act (Regulation 2024/1689). This skill integrates the GDPR Art. 35 DPIA framework with AI-specific risk assessment, encompassing training data lawfulness, Art. 22 automated decision-making implications, algorithmic fairness, and the NIST AI Risk Management Framework MAP function. The assessment methodology draws from the EDPB-EDPS Joint Opinion 5/2021 on the AI Act proposal and subsequent EDPB Guidelines 06/2025 on AI and data protection.
Legal Framework
GDPR Provisions Applicable to AI
| Provision | Application to AI Systems |
|---|---|
| Art. 5(1)(a) — Lawfulness, fairness, transparency | AI processing must have a lawful basis; the logic of AI decisions must be explainable to data subjects |
| Art. 5(1)(b) — Purpose limitation | Training data collected for one purpose cannot be used to train AI models for an incompatible purpose without further lawful basis |
| Art. 5(1)(c) — Data minimisation | AI models should not require more personal data than necessary; synthetic data and anonymisation should be considered |
| Art. 5(1)(d) — Accuracy | AI outputs affecting individuals must be accurate; model drift must be monitored |
| Art. 6(1) — Lawful basis | Each stage of AI processing (data collection, model training, inference, output use) requires a lawful basis |
| Art. 9 — Special categories | Training on health, biometric, genetic, racial, political, religious, sexual orientation, or trade union data requires an Art. 9(2) exemption |
| Art. 13-14 — Transparency | Privacy notices must disclose the existence of automated decision-making, meaningful information about the logic involved, and the significance and envisaged consequences |
| Art. 22 — Automated decision-making | Data subjects have the right not to be subject to decisions based solely on automated processing that produce legal effects or similarly significantly affect them, with exceptions under Art. 22(2) |
| Art. 25 — Data protection by design | AI systems must embed privacy protections from the design phase: privacy-preserving ML techniques, differential privacy, federated learning |
| Art. 35 — DPIA | AI systems meeting EDPB WP248rev.01 criteria (evaluation/scoring, automated decision-making, innovative technology) require a DPIA |
EU AI Act (Regulation 2024/1689) Risk Classification
| Risk Category | Description | AI Act Requirements | GDPR Overlap |
|---|---|---|---|
| Unacceptable Risk (Art. 5) | AI practices prohibited outright: social scoring by public authorities, real-time remote biometric identification in public spaces (with exceptions), emotion recognition in workplace/education, untargeted scraping for facial recognition databases | Prohibited — cannot be deployed | Art. 9 special categories, Art. 35 DPIA |
| High Risk (Annex III) | AI systems in listed areas: biometrics, critical infrastructure, education, employment, essential services, law enforcement, migration, administration of justice | Conformity assessment, risk management system, data governance, transparency, human oversight, accuracy/robustness/cybersecurity | Art. 22, Art. 35 DPIA, Art. 25 DPbD |
| Limited Risk (Art. 50) | AI systems with specific transparency obligations: chatbots, emotion recognition, deep fakes | Transparency obligations (notify users they are interacting with AI) | Art. 13-14 transparency |
| Minimal Risk | All other AI systems | No specific AI Act obligations; voluntary codes of practice | Standard GDPR compliance |
EDPB-EDPS Joint Opinion 5/2021
Key recommendations from the Joint Opinion on the AI Act proposal:
- Prohibition of remote biometric identification in public spaces — supported without exceptions
- AI-specific DPIA — the EDPB recommended mandatory DPIAs for all high-risk AI systems, not limited to Art. 35 triggers
- Right to explanation — Art. 22(3) right to obtain human intervention and contest decisions should apply to all AI-assisted decisions, not only fully automated ones
- Training data governance — controllers must ensure lawfulness of personal data used for AI training throughout the model lifecycle, including when training data is subsequently deleted
- Purpose limitation for AI training — repurposing personal data for AI model training requires a compatibility assessment under Art. 6(4) or a separate lawful basis
Assessment Methodology
Phase 1: AI System Classification (Week 1)
1.1 AI Act Risk Classification
Determine the AI system's risk category under the AI Act:
- Check Art. 5 prohibited practices list — if the system falls within a prohibition, it cannot be deployed.
- Check Annex III high-risk categories:
- Biometric identification and categorisation of natural persons
- Management and operation of critical infrastructure
- Education and vocational training (admissions, assessment, proctoring)
- Employment, worker management, and access to self-employment (recruitment, task allocation, performance evaluation, termination)
- Access to and enjoyment of essential private and public services (credit scoring, insurance pricing, emergency services dispatch)
- Law enforcement (crime risk assessment, polygraph, evidence evaluation)
- Migration, asylum, and border control
- Administration of justice and democratic processes
- Check Art. 50 transparency obligations for limited-risk AI.
- If not high-risk or limited-risk, classify as minimal risk.
1.2 GDPR Processing Assessment
- Identify all personal data processed at each AI lifecycle stage: data collection, data preparation, model training, model validation, inference/deployment, output use.
- For each stage, identify the lawful basis under Art. 6(1).
- If special category data is involved, identify the Art. 9(2) condition.
- Assess whether Art. 22 applies (solely automated decisions with legal or significant effect).
Phase 2: Training Data Lawfulness (Week 2)
2.1 Data Collection Lawfulness
| Assessment Question | Requirement |
|---|---|
| Was the training data collected with a lawful basis under Art. 6(1)? | Each data source must have an identified lawful basis |
| Is the use of data for AI training compatible with the original collection purpose? | Art. 6(4) compatibility assessment or Art. 5(1)(b) further processing analysis |
| Was consent obtained for AI training specifically? | If relying on Art. 6(1)(a), consent must be specific, informed, and freely given for the AI training purpose |
| If using legitimate interest, has a balancing test been conducted? | Art. 6(1)(f) requires documented LIA including AI-specific impacts |
| Does training data include special categories? | Art. 9(2) exemption required; Art. 9(2)(j) scientific research may apply with safeguards |
2.2 Data Quality and Bias Assessment
| Assessment Area | Requirements |
|---|---|
| Representativeness | Training data must be representative of the population the AI system will be applied to. Underrepresentation of demographic groups must be identified and addressed. |
| Label accuracy | If supervised learning, labels must be accurate and free from historical bias. Human labellers must be trained on anti-discrimination principles. |
| Temporal validity | Training data must reflect current conditions. Stale training data can produce discriminatory outputs. |
| Proxy variables | Identify features that may serve as proxies for protected characteristics (postcode as proxy for ethnicity, name as proxy for gender). |
| Data provenance | Document the source, collection methodology, and processing history for all training data. |
2.3 Training Data Retention
- Art. 5(1)(e) storage limitation applies to training data.
- Once training is complete, assess whether original training data must be retained or can be deleted.
- If training data is deleted, document how data subject rights (access, erasure) will be facilitated regarding data embedded in the trained model.
- Consider whether the trained model itself constitutes personal data (if individual data points can be extracted through model inversion attacks).
Phase 3: Art. 22 Automated Decision-Making Assessment (Week 3)
3.1 Art. 22(1) Applicability Test
Does the AI system make decisions about individuals?
├─ NO → Art. 22 does not apply.
└─ YES → Continue.
│
Are decisions based solely on automated processing?
├─ NO → Art. 22(1) does not apply, but Art. 13-14 transparency still applies.
│ (Note: "meaningful human involvement" must be genuine, not rubber-stamping.)
└─ YES → Continue.
│
Do decisions produce legal effects or similarly significantly affect the individual?
├─ NO → Art. 22(1) does not apply.
└─ YES → Art. 22(1) applies. The individual has the right not to be subject
to the decision unless an Art. 22(2) exception applies.
3.2 Art. 22(2) Exceptions
| Exception | Requirements |
|---|---|
| Art. 22(2)(a) — Necessary for contract | Decision must be necessary for entering into or performing a contract with the data subject |
| Art. 22(2)(b) — Authorised by law | Union or Member State law must authorise the decision and provide suitable safeguards |
| Art. 22(2)(c) — Explicit consent | Data subject has given explicit consent to the automated decision |
3.3 Art. 22(3) Safeguards
Even where an Art. 22(2) exception applies, the controller must implement suitable measures including:
- Right to obtain human intervention
- Right to express the data subject's point of view
- Right to contest the decision
- Meaningful information about the logic involved (not the full algorithm, but sufficient to understand the factors and their relative weight)
Phase 4: Algorithmic Bias and Fairness Assessment (Week 3-4)
4.1 Fairness Metrics
| Metric | Description | Application |
|---|---|---|
| Demographic parity | Positive outcome rates should be equal across protected groups | Credit scoring, hiring |
| Equalized odds | True positive and false positive rates should be equal across groups | Criminal risk assessment, fraud detection |
| Predictive parity | Positive predictive value should be equal across groups | Medical diagnosis, recidivism prediction |
| Individual fairness | Similar individuals should receive similar outcomes | Loan pricing, insurance premium calculation |
| Counterfactual fairness | Outcome should not change if only the protected characteristic changes | Any decision-making system |
4.2 Bias Testing Protocol
- Define protected characteristics relevant to the jurisdiction and use case (race, gender, age, disability, religion, sexual orientation per Equality Act 2010 / EU Charter Art. 21).
- Partition test dataset by protected characteristics.
- Calculate fairness metrics for each group.
- Identify disparate impact: if the selection rate for any protected group is less than 80% of the rate for the group with the highest rate (four-fifths rule from US EEOC, widely adopted as a starting benchmark).
- If disparate impact is detected, assess whether the disparity is justified by a legitimate, non-discriminatory factor.
- Document bias testing methodology, results, and remediation actions.
Phase 5: NIST AI RMF MAP Function Integration (Week 4)
The NIST AI Risk Management Framework (AI RMF 1.0, January 2023) MAP function identifies the context, capabilities, and potential impacts of AI systems. Integrate the following MAP subcategories:
| Subcategory | Assessment Action |
|---|---|
| MAP 1.1 | Document the intended purpose, context of use, and deployment environment |
| MAP 1.2 | Document interdependent and interconnected systems |
| MAP 1.5 | Identify intended users, affected individuals, and stakeholders |
| MAP 1.6 | Assess impacts on individuals, groups, communities, organisations, and society |
| MAP 2.1 | Establish the AI system's knowledge limits and conditions where it may fail |
| MAP 2.2 | Document scientific integrity of training and testing methodologies |
| MAP 2.3 | Assess environmental impact of AI system training and deployment |
| MAP 3.1 | Document potential benefits of the AI system |
| MAP 3.2 | Document potential costs, risks, and negative impacts |
| MAP 3.4 | Map risks specifically to affected communities and stakeholders |
| MAP 3.5 | Document likelihood and severity of identified risks |
| MAP 5.1 | Engage with diverse stakeholders and affected communities |
| MAP 5.2 | Engage with domain experts, AI practitioners, and sociotechnical experts |
Phase 6: Combined Risk Assessment and Reporting (Week 5-6)
- Consolidate GDPR DPIA risks, AI Act compliance gaps, and NIST AI RMF findings.
- For each risk, assess likelihood and severity using the DPIA risk matrix.
- Identify mitigation measures combining privacy controls and AI governance controls.
- Calculate residual risk levels.
- If high-risk AI under the AI Act, prepare conformity assessment documentation.
- Submit to DPO for Art. 35(2) advice.
- If residual risk remains high, initiate Art. 36 prior consultation.
AI-Specific Privacy Risks
| Risk Category | Description | Mitigation Approach |
|---|---|---|
| Model inversion | Attacker reconstructs training data from model outputs | Differential privacy during training, output perturbation, access controls on model API |
| Membership inference | Attacker determines whether a specific individual's data was in the training set | Regularisation, differential privacy, limiting model confidence scores |
| Data poisoning | Malicious manipulation of training data to bias model outputs | Training data provenance verification, anomaly detection, robust training techniques |
| Concept drift | Model accuracy degrades over time as real-world data distribution changes | Continuous monitoring, automated retraining triggers, human review of edge cases |
| Explanation manipulation | Gaming of AI explanations to hide discriminatory factors | Multiple explanation methods, adversarial testing of explanation consistency |
| Feedback loops | AI decisions create data that reinforces existing biases | Regular bias auditing, human-in-the-loop review, outcome monitoring by demographic group |
Common Assessment Deficiencies
- Treating AI training as a single processing operation: Each stage (collection, preparation, training, inference) is a distinct processing operation requiring its own lawful basis assessment.
- Relying on anonymisation claims without verification: Data claimed to be anonymised may be re-identifiable through model inversion or linkage attacks.
- Inadequate Art. 22 human involvement: Rubber-stamp human review of AI decisions does not constitute meaningful human involvement.
- Ignoring purpose limitation for training data: Using customer data collected for service delivery to train AI models without compatibility assessment.
- No bias testing on protected characteristics: Fairness assessment limited to overall accuracy without demographic disaggregation.
- Static assessment: AI systems evolve through retraining; the assessment must be updated when the model is retrained or the deployment context changes.
Enforcement Precedents
- Italian Garante vs Clearview AI (2022): EUR 20 million fine for scraping biometric data from the internet to train facial recognition AI without lawful basis, consent, or DPIA.
- Italian Garante vs Deliveroo (2021): EUR 2.5 million fine for algorithmic management of food delivery riders constituting Art. 22 automated decision-making without adequate safeguards. The algorithm assigned reputation scores and allocated delivery shifts without meaningful human review.
- Hungarian DPA vs Budapest Bank (2023): Fine for automated credit scoring system that used postcode as a proxy for ethnicity without bias impact assessment.
- CNIL vs Clearview AI (2022): EUR 20 million fine; CNIL ordered deletion of all data of French residents and prohibited further collection.
- Dutch DPA vs Tax Authority (2020): EUR 3.7 million fine for algorithmic profiling system (SyRI) that disproportionately targeted ethnic minorities in fraud detection, violating ECHR Art. 8 and Art. 14.
Assessing Biometric Processing Privacy
Overview
Biometric data is classified as a special category of personal data under GDPR Art. 9(1) when processed for the purpose of uniquely identifying a natural person. Processing biometric data on a large scale triggers a mandatory DPIA under Art. 35(3)(b). This skill provides a comprehensive DPIA methodology for biometric systems including facial recognition, fingerprint identification, voice recognition, iris scanning, vein pattern analysis, and behavioural biometrics (gait, typing patterns, signature dynamics).
Legal Framework for Biometric Data
GDPR Definition — Art. 4(14)
"'Biometric data' means personal data resulting from specific technical processing relating to the physical, physiological or behavioural characteristics of a natural person, which allow or confirm the unique identification of that natural person, such as facial images or dactyloscopic data."
Art. 9(1) — Prohibition on Processing Special Categories
Processing of biometric data for the purpose of uniquely identifying a natural person is prohibited unless one of the Art. 9(2) exemptions applies.
Critical distinction: Art. 9 only applies when biometric data is processed "for the purpose of uniquely identifying" a person. A photograph used for illustration purposes is not Art. 9 data; the same photograph processed through facial recognition software to identify the person is Art. 9 data.
Art. 9(2) Exemptions Applicable to Biometric Processing
| Exemption | Reference | Application to Biometrics |
|---|---|---|
| Explicit consent | Art. 9(2)(a) | Employee consent often not freely given due to power imbalance (WP29 Opinion 2/2017). Consumer biometric consent must meet Art. 7 standards. |
| Employment, social security, social protection law | Art. 9(2)(b) | Member State law may authorise biometric processing in the employment context (e.g., biometric access control for high-security areas). |
| Vital interests | Art. 9(2)(c) | Limited to emergency situations where biometric identification is needed to protect someone's life. |
| Substantial public interest | Art. 9(2)(g) | Member State law basis required. May apply to law enforcement biometrics where authorised by specific legislation. |
| Health or social care | Art. 9(2)(h) | Biometric patient identification in healthcare settings. |
| Public health | Art. 9(2)(i) | Biometric contact tracing during health emergencies (subject to proportionality). |
| Archiving, scientific research, statistics | Art. 9(2)(j) | Biometric research (e.g., medical imaging analysis) with Art. 89(1) safeguards. |
Art. 35(3)(b) — Mandatory DPIA Trigger
Processing on a large scale of special categories of data referred to in Art. 9(1), including biometric data processed for unique identification, requires a DPIA. "Large scale" factors per WP248rev.01:
- Number of data subjects (in absolute terms or as a proportion of the relevant population)
- Volume of data and/or range of data items
- Duration or permanence of the processing
- Geographic scope
EDPB Guidelines 3/2019 on Video Surveillance
Key provisions relevant to facial recognition CCTV:
- Facial recognition in public spaces generally constitutes systematic monitoring of publicly accessible areas (Art. 35(3)(c)) in addition to large-scale biometric processing (Art. 35(3)(b)).
- Facial recognition for access control is less intrusive than identification in public spaces but still requires DPIA.
- Purpose limitation: biometric templates created for access control must not be repurposed for attendance monitoring or performance management.
- Storage limitation: biometric templates should be stored on a device controlled by the data subject (e.g., access card) rather than a central database where possible.
Biometric System Types and Risk Assessment
Verification (1:1 Matching)
One-to-one comparison of a live biometric sample against a stored template for the claimed identity. Used for access control, device unlock, payment authentication.
| Risk Factor | Assessment |
|---|---|
| Data subjects | Defined, enrolled individuals |
| Volume of data | Limited to enrolled population |
| Proportionality | Generally more proportionate than identification |
| Storage recommendation | Template stored on user's device or card (decentralised) |
| Art. 35(3)(b) trigger | Depends on scale of enrolled population |
Identification (1:N Matching)
One-to-many comparison of a live biometric sample against a database of templates to determine identity. Used for law enforcement, border control, surveillance.
| Risk Factor | Assessment |
|---|---|
| Data subjects | Potentially unlimited (all persons in the capture area) |
| Volume of data | Can be very large (entire population databases) |
| Proportionality | Highly intrusive; requires strong justification |
| Storage recommendation | Centralised database is typically required for 1:N matching |
| Art. 35(3)(b) trigger | Almost always triggered |
Categorisation
Classification of individuals into groups based on biometric characteristics (age, gender, ethnicity, emotion) without uniquely identifying them. Used for analytics, targeted advertising, audience measurement.
| Risk Factor | Assessment |
|---|---|
| Art. 9 applicability | May not fall under Art. 9 if not used for unique identification, but still high risk |
| AI Act classification | Emotion recognition in workplace/education prohibited (Art. 5(1)(f)) |
| Discrimination risk | Categorisation by race, ethnicity, or emotion raises equality law concerns |
DPIA Content for Biometric Systems
Systematic Description
For biometric systems, the Art. 35(7)(a) systematic description must include:
| Element | Required Detail |
|---|---|
| Biometric modality | Facial, fingerprint, iris, voice, vein, gait, or multi-modal |
| Processing mode | Verification (1:1) or identification (1:N) or categorisation |
| Capture environment | Controlled (sensor/scanner) or uncontrolled (CCTV, ambient camera) |
| Template storage | Centralised database, decentralised (user device/card), or encrypted enclave |
| Template format | Proprietary template, ISO 19794 standard, or raw biometric data |
| Matching algorithm | Vendor algorithm (specify), open-source algorithm, or custom development |
| Accuracy metrics | False Acceptance Rate (FAR), False Rejection Rate (FRR), Equal Error Rate (EER) |
| Liveness detection | Anti-spoofing measures (presentation attack detection) |
| Fallback mechanism | Alternative identification method when biometric fails |
| Retention period | Template retention, raw biometric data retention, audit log retention |
Necessity and Proportionality for Biometric Processing
The proportionality assessment for biometric systems must be rigorous because biometric data:
- Is permanent (unlike passwords, biometric characteristics cannot be changed if compromised)
- Is uniquely identifying (biometric data is inherently linked to the individual)
- Creates heightened risk of function creep (biometric templates can be repurposed)
- Has disproportionate impact if breached (biometric data cannot be reissued)
| Proportionality Question | Assessment Standard |
|---|---|
| Is biometric processing necessary, or can a non-biometric alternative achieve the same purpose? | Badge/card access, PIN, password, or multi-factor authentication without biometrics |
| Is the biometric modality the least intrusive option? | Fingerprint is generally less intrusive than facial recognition; on-device verification less intrusive than centralised identification |
| Is the scale of biometric processing proportionate? | Processing all persons in an area (identification) is less proportionate than processing enrolled volunteers (verification) |
| Is the retention of biometric data minimised? | On-card template storage preferred over centralised database; raw biometric images should not be retained after template extraction |
Risk Register for Biometric DPIA
Common Biometric Processing Risks
| Risk ID | Risk | Likelihood | Severity | Typical Level |
|---|---|---|---|---|
| BIO-R1 | Biometric data breach — templates or raw data exposed to unauthorised parties | Possible | Maximum | Very High |
| BIO-R2 | Function creep — biometric data collected for access control repurposed for surveillance or attendance monitoring | Likely | Significant | High |
| BIO-R3 | Discriminatory accuracy — facial recognition performs worse on certain demographic groups (skin colour, age, gender) | Likely | Significant | High |
| BIO-R4 | False rejection denying legitimate access — disabled individuals, elderly, or those with skin conditions experience higher rejection rates | Possible | Significant | High |
| BIO-R5 | Spoofing or presentation attacks — fraudulent biometric samples (photos, masks, artificial fingerprints) bypass security | Possible | Significant | High |
| BIO-R6 | Chilling effect — knowledge of biometric surveillance alters behaviour in public or workplace spaces | Likely | Limited | High |
| BIO-R7 | Irreversibility — unlike passwords, compromised biometric data cannot be changed or reissued | Almost certain | Maximum | Very High |
| BIO-R8 | Third-party capture — biometric data of non-enrolled individuals incidentally captured by the system | Likely | Limited | High |
Mitigation Measures for Biometric Processing
| Measure | Type | Risk Addressed |
|---|---|---|
| On-device or on-card template storage (no centralised database) | Technical (DPbD) | BIO-R1, BIO-R7 |
| Template protection: cancellable biometrics or biometric encryption (BioHashing, fuzzy vault) | Technical | BIO-R1, BIO-R7 |
| Liveness detection / presentation attack detection (ISO 30107 compliance) | Technical | BIO-R5 |
| Purpose limitation enforcement through technical access controls | Technical + Organisational | BIO-R2 |
| Demographic accuracy testing across skin colour, age, gender groups (NIST FRVT benchmarks) | Technical | BIO-R3 |
| Non-biometric fallback mechanism (PIN, card, helpdesk override) | Organisational | BIO-R4 |
| Signage and transparency notices in capture areas | Organisational (Transparency) | BIO-R6, BIO-R8 |
| Automatic deletion of raw biometric images after template extraction | Technical (Data minimisation) | BIO-R1 |
| Regular penetration testing of biometric system | Technical (Security) | BIO-R5 |
Enforcement Precedents
- CNIL vs Clearview AI (2022): EUR 20 million fine for mass collection of facial images from internet for biometric identification database without consent, transparency, or DPIA.
- ICO vs Clearview AI (2022): GBP 7.5 million fine for same biometric processing; enforcement notice ordering deletion of UK residents' biometric data.
- Swedish DPA vs Skelleftea Municipality (2019): SEK 200,000 fine for school using facial recognition for student attendance tracking. Consent relied upon was not freely given due to power imbalance. Less intrusive alternatives were available.
- ICO vs Serco Leisure (2022): Enforcement notice for requiring employees to use facial recognition for time and attendance at leisure centres. No DPIA conducted; no less intrusive alternative offered.
- CNIL vs Clearview AI (2023): EUR 5.2 million additional penalty for non-compliance with the 2022 order, demonstrating the consequences of continued biometric processing in violation of enforcement.
- French Conseil d'Etat vs Presto (2020): Upheld CNIL enforcement against biometric time-and-attendance system for employees; found that fingerprint scanning for attendance was disproportionate when badge systems were available.
- Italian Garante vs P&G (Gillette) (2022): Enforcement action against automated age estimation system using facial analysis in retail stores without DPIA and without valid consent.
Assessing Cloud Migration Privacy
Overview
Migrating personal data to cloud infrastructure (IaaS, PaaS, SaaS) introduces fundamental changes to the data protection landscape: the controller cedes physical control over personal data to a cloud service provider (CSP) acting as processor, creating dependencies on the CSP's security measures, geographic infrastructure, and sub-processor chain. This skill provides a DPIA methodology covering the Art. 28 controller-processor relationship, international transfer assessment under Chapter V, encryption and key management requirements, the shared responsibility model, and sub-processor governance.
Legal Framework
GDPR Art. 28 — Processor Obligations
Art. 28(1) requires the controller to use only processors providing sufficient guarantees to implement appropriate technical and organisational measures to ensure processing meets GDPR requirements.
Art. 28(3) mandates a binding contract or legal act setting out:
- Subject matter and duration of processing
- Nature and purpose of processing
- Type of personal data and categories of data subjects
- Obligations and rights of the controller
- Specific processor obligations including: processing only on documented instructions, confidentiality, security measures, sub-processor approval, data subject rights assistance, deletion/return on termination, audits and inspections
Art. 28(2) — Sub-Processor Authorisation
The processor shall not engage another processor without prior specific or general written authorisation of the controller. In the case of general written authorisation, the processor shall inform the controller of any intended changes concerning the addition or replacement of sub-processors, giving the controller the opportunity to object.
Cloud-Specific Regulatory Guidance
| Authority | Guidance | Key Requirements |
|---|---|---|
| EDPB | Guidelines 07/2020 on controller and processor concepts | Clarified CSP role determination: CSPs acting strictly under controller instructions are processors; CSPs determining purposes of processing for their own use are (joint) controllers |
| CNIL | Recommendations on cloud computing (2022) | Recommended DPIA for cloud migrations involving sensitive data; encryption with customer-held keys; exit strategy requirements |
| BSI (Germany) | C5 Cloud Computing Compliance Catalogue | Technical security criteria for cloud services used by German public sector and recommended for private sector |
| ENISA | Cloud Security Guide for SMEs (2015, updated 2023) | Risk assessment framework for cloud adoption including data protection considerations |
Cloud Service Models and Privacy Responsibilities
Shared Responsibility Model
| Responsibility | IaaS | PaaS | SaaS |
|---|---|---|---|
| Physical infrastructure security | CSP | CSP | CSP |
| Network security | Shared | CSP | CSP |
| Operating system | Customer | CSP | CSP |
| Application security | Customer | Shared | CSP |
| Data classification | Customer | Customer | Customer |
| Data encryption (at rest) | Customer | Shared | CSP (configuration by customer) |
| Data encryption (in transit) | Customer | Shared | CSP |
| Identity and access management | Customer | Shared | Shared |
| Data backup and recovery | Customer | Shared | CSP (configuration by customer) |
| Logging and monitoring | Customer | Shared | CSP (access by customer) |
| Compliance and governance | Customer | Customer | Customer |
| Data subject rights facilitation | Customer | Customer | Customer |
| DPIA | Customer | Customer | Customer |
| Data processing agreement | Customer | Customer | Customer |
Key Privacy Considerations by Service Model
IaaS (Infrastructure as a Service)
- Controller retains most control but also most responsibility
- CSP has physical access to infrastructure but not application-layer data if encrypted
- Customer responsible for OS patching, application security, encryption implementation
- Data residency controlled by customer's region selection
PaaS (Platform as a Service)
- CSP manages infrastructure and platform; customer manages application and data
- CSP may have access to data through platform diagnostics and logging
- Sub-processor chain may be extensive (platform dependencies)
- Data portability may be limited by platform lock-in
SaaS (Software as a Service)
- CSP has broadest access to personal data (application-level access)
- Customer has least control over security measures
- Sub-processor chain often complex and opaque
- Data portability and exit strategy are critical concerns
DPIA Methodology for Cloud Migration
Phase 1: Cloud Processing Mapping (Week 1)
- Inventory all personal data to be migrated to cloud.
- Classify data by sensitivity: public, internal, confidential, restricted.
- Map data flows: on-premises to cloud, cloud to cloud, cloud to third parties.
- Identify all CSP services that will process personal data.
- Determine CSP data processing locations (regions, zones, edge locations).
- Identify the complete sub-processor chain for each CSP service.
Phase 2: Controller-Processor Analysis (Week 2)
- Determine the CSP's role for each service:
- Processor: CSP processes data only on controller's documented instructions
- Joint controller: CSP determines purposes of processing (e.g., CSP uses data for service improvement, analytics, AI training)
- Review the CSP's Data Processing Agreement (DPA) for Art. 28(3) compliance.
- Verify the CSP's sub-processor list and notification mechanism.
- Assess the CSP's audit provisions: on-site audit rights, SOC 2 reports, ISO 27001 certification, penetration test results.
- Review data return and deletion provisions for contract termination.
Phase 3: International Transfer Assessment (Week 3)
- Identify all CSP data processing locations.
- For each location outside the EEA:
- Check for an adequacy decision (Art. 45)
- Verify SCCs are in place (Art. 46(2)(c))
- Conduct Transfer Impact Assessment per EDPB Recommendations 01/2020
- Assess sub-processor locations for international transfers.
- Evaluate technical supplementary measures (encryption with customer-held keys).
- Document the transfer mechanism for each third-country transfer.
Phase 4: Security and Encryption Assessment (Week 3-4)
- Evaluate CSP encryption capabilities:
- Encryption at rest (algorithm, key length, key management)
- Encryption in transit (TLS version, cipher suites)
- Customer-managed encryption keys (BYOK/HYOK) availability
- Client-side encryption capability
- Assess access controls:
- CSP staff access to customer data (break-glass procedures, access logging)
- Customer IAM integration (SSO, MFA, RBAC)
- Privileged access management
- Evaluate data isolation:
- Multi-tenant vs dedicated infrastructure
- Logical isolation controls
- Data commingling risks
- Review incident response:
- CSP breach notification timeline (Art. 33(2) requires processor to notify controller without undue delay)
- CSP incident response capabilities
- Customer access to security event logs
Phase 5: Risk Assessment (Week 4-5)
| Risk | Description | Typical Level |
|---|---|---|
| CR-R1 | CSP data breach exposing customer personal data | High |
| CR-R2 | Government access to data via CSP (FISA 702, CLOUD Act for US CSPs) | Medium-High |
| CR-R3 | Sub-processor change introducing new risks without controller awareness | Medium |
| CR-R4 | Vendor lock-in preventing data portability or migration to alternative provider | Medium |
| CR-R5 | CSP service termination or insolvency | Medium |
| CR-R6 | CSP using personal data for own purposes (service improvement, AI training) | Medium |
| CR-R7 | Insufficient logging preventing incident investigation | Medium |
| CR-R8 | Data residency violation through CSP infrastructure changes | Medium |
Phase 6: Mitigation and Approval (Week 5-6)
- Implement mitigations for each identified risk.
- Negotiate DPA amendments where CSP standard terms are insufficient.
- Implement customer-managed encryption keys where feasible.
- Establish CSP monitoring procedures (certification renewals, sub-processor changes, incident notifications).
- DPO review and approval.
- Schedule review: annually or upon CSP service change, new sub-processor, or security incident.
Common Cloud Migration DPIA Deficiencies
- Accepting CSP DPA without review: CSP standard DPAs may not meet all Art. 28(3) requirements or may include controller-unfavourable terms (e.g., broad sub-processor general authorisation without objection right).
- Ignoring sub-processor chain: CSPs may use dozens of sub-processors; the controller remains responsible for ensuring adequate protection at each level.
- Relying on CSP-managed encryption only: CSP-managed keys provide protection against physical theft but not against government compulsion to the CSP or CSP staff access.
- No exit strategy: Failure to plan for data return and deletion on contract termination, creating lock-in risk and residual data risk.
- No transfer assessment for multi-region CSP: CSP may replicate data across regions including non-EEA locations for redundancy, backup, or CDN purposes.
Enforcement Precedents
- DPC Ireland vs Meta (2023): EUR 1.2 billion fine for transferring EU user data to US-based cloud infrastructure without valid transfer mechanism.
- DPC Ireland vs Microsoft (LinkedIn, 2023): EUR 310 million fine related to advertising data processing; highlighted the importance of controller-processor role clarity in cloud services.
- CNIL vs Microsoft 365 Education (2022): Raised concerns about data transfers to the US through Microsoft cloud services and insufficient DPA provisions.
- German Federal Data Protection Commissioner (2022): Advised against use of Microsoft 365 in federal agencies due to insufficient transparency about data processing and transfers.
- Dutch DPA DPIA on Microsoft Office 365 (SLM Rijk, 2022): Identified eight high-risk areas in Microsoft 365 processing including diagnostic data collection, data transfers, and insufficient controller access to processing records.
Comparing PIA Methodologies
Overview
Multiple established methodologies exist for conducting Privacy Impact Assessments: the CNIL PIA tool, ICO DPIA template, NIST Privacy Framework, and ISO/IEC 29134:2017. Each methodology reflects its originating regulatory context, organisational assumptions, and privacy philosophy. Selecting the appropriate methodology — or combining elements from several — is critical for producing assessments that satisfy regulatory expectations, align with organisational maturity, and address the actual risks of the processing activity. This skill provides a structured comparison framework for methodology selection.
Methodology Profiles
1. CNIL PIA Tool (France)
Origin: Commission Nationale de l'Informatique et des Libertes (CNIL), first published 2015, updated 2018 for GDPR alignment.
Structure:
- Step 1: Context — Describe the processing, its purposes, the data processed, and the actors involved.
- Step 2: Fundamental Principles — Assess compliance with necessity, proportionality, data subject rights, and obligations (Art. 5, 6, 9, 12-22, 28, 44).
- Step 3: Risks — Identify feared events (illegitimate access, unwanted modification, disappearance of data), assess severity and likelihood.
- Step 4: Validation — Map risks against controls. Decide whether to accept residual risk, implement additional measures, or consult the supervisory authority.
Key Features:
- Three feared events model (access, modification, disappearance) structured around CIA triad adapted for privacy.
- Severity scale: Negligible, Limited, Significant, Maximum.
- Likelihood scale: Negligible, Limited, Significant, Maximum.
- Risk matrix: 4x4 grid mapping severity against likelihood.
- Explicit link to Art. 35 and Art. 36 prior consultation threshold.
- Open-source PIA software tool available for download.
Regulatory Acceptance:
- Required format for French supervisory authority submissions.
- Accepted by Belgian DPA (APD/GBA) as compliant methodology.
- Referenced by EDPB as an example of good practice in WP248rev.01.
- Widely used across francophone jurisdictions (Belgium, Luxembourg, Switzerland).
2. ICO DPIA Template (United Kingdom)
Origin: Information Commissioner's Office (ICO), UK. Published as part of ICO GDPR guidance, updated for UK GDPR post-Brexit.
Structure:
- Step 1: Identify the need for a DPIA — Screening checklist against Art. 35(3) triggers and ICO's published list of processing requiring DPIA.
- Step 2: Describe the processing — Nature, scope, context, purposes. Data flows including recipients and transfers.
- Step 3: Consultation process — Record of consultation with data subjects, DPO, and other stakeholders.
- Step 4: Assess necessity and proportionality — Lawful basis, purpose limitation, data minimisation, accuracy, storage limitation, security, international transfers.
- Step 5: Identify and assess risks — Risks to individuals organised by source, nature of harm, severity, likelihood.
- Step 6: Identify measures to mitigate risks — For each risk, identify measures that reduce it to an acceptable level.
- Step 7: Sign off and record outcomes — DPO advice, controller decision, integration with processing register.
Key Features:
- Seven-step linear process designed for practical completion.
- Emphasis on consultation with data subjects (Step 3) — a distinctive feature not prominent in other methodologies.
- Risk assessment focused on harm to individuals rather than CIA triad.
- Explicit requirement to record DPO advice and whether it was followed.
- Template available as downloadable Word document.
- Screening checklist integrated into Step 1.
Regulatory Acceptance:
- Required format for UK ICO submissions and prior consultation.
- Accepted by Irish DPC as compliant methodology for UK-Ireland processing.
- Used as basis for several Member State DPA templates (Denmark, Netherlands).
- Recommended by UK government for public sector processing.
3. NIST Privacy Framework (United States)
Origin: National Institute of Standards and Technology (NIST), Version 1.0 published January 2020. Voluntary framework.
Structure:
- Core: Five functions — IDENTIFY, GOVERN, CONTROL, COMMUNICATE, PROTECT.
- Profiles: Current state and target state assessment for each subcategory.
- Implementation Tiers: Tier 1 (Partial) to Tier 4 (Adaptive) maturity levels.
Key Features:
- Not a PIA methodology per se, but provides the organisational context within which PIAs are conducted.
- IDENTIFY function (ID.RA) maps directly to risk assessment activities in PIA.
- Complementary to NIST Cybersecurity Framework (CSF) — designed for joint deployment.
- Sector-agnostic: applicable across all industries, not limited to specific regulatory context.
- Privacy Engineering objectives: Predictability, Manageability, Disassociability.
- No prescribed risk scale or matrix — organisations define their own.
Regulatory Acceptance:
- Not a regulatory requirement but widely referenced by US federal agencies (FTC, HHS, DOE).
- Accepted as evidence of privacy program maturity by US state regulators (CCPA/CPRA, VCDPA, CPA).
- Referenced by APEC CBPR system for cross-border data flows.
- Increasingly referenced by EU organisations as complementary to GDPR DPIA.
4. ISO/IEC 29134:2017 — Privacy Impact Assessment Guidelines
Origin: International Organization for Standardization, published 2017. International standard.
Structure:
- Clause 6: PIA preparation — Determine necessity, establish PIA team, develop PIA plan, stakeholder engagement.
- Clause 7: PIA execution — Information gathering, data flow analysis, privacy risk analysis (identification, estimation, evaluation), privacy risk treatment.
- Clause 8: PIA follow-up — PIA report, publication of summary, implementation of treatment plan, audit/review.
Key Features:
- Most comprehensive and structured methodology — designed as auditable standard.
- Risk assessment based on ISO 31000 risk management framework.
- Privacy risk = likelihood x consequence (consequence to the data subject, not the organisation).
- Requires formal PIA plan before assessment begins.
- Mandates stakeholder engagement including data subjects.
- PIA report has prescribed structure (Clause 8.2) with 13 required sections.
- Integrates with ISO/IEC 27701 (privacy information management) and ISO/IEC 27001 (information security).
Regulatory Acceptance:
- Accepted by supervisory authorities globally as evidence of systematic approach.
- Referenced by EDPB in WP248rev.01 as an example of PIA methodology.
- Required or recommended by several Asian-Pacific DPAs (Singapore PDPC, South Korea PIPC, Japan PPC).
- Certification bodies use ISO 29134 as assessment framework for Art. 42 GDPR certification schemes.
Methodology Comparison Matrix
| Dimension | CNIL PIA | ICO DPIA | NIST PF | ISO 29134 |
|---|---|---|---|---|
| Regulatory origin | French DPA (CNIL) | UK DPA (ICO) | US NIST | International (ISO/IEC) |
| Legal framework | GDPR | UK GDPR | Sector-agnostic | International |
| Risk model | 3 feared events (CIA-adapted) | Harm to individuals | Organisation-defined | ISO 31000-based |
| Risk scale | 4x4 (Negligible to Maximum) | Qualitative (Low/Medium/High) | Organisation-defined tiers | Likelihood x Consequence |
| Steps/phases | 4 steps | 7 steps | 5 functions | 3 clauses (prep/execute/follow-up) |
| Data subject consultation | Recommended | Explicitly required (Step 3) | COMMUNICATE function | Required (Clause 6.4) |
| DPO involvement | Required | Required with advice recording | N/A (no DPO concept) | Recommended |
| Tool availability | Open-source software | Word template | Excel self-assessment | No official tool |
| Cost | Free | Free | Free | Standard purchase required (~CHF 166) |
| Certification alignment | None | None | NIST CSF alignment | ISO 27701, ISO 27001 |
| Typical completion time | 2-4 weeks | 1-3 weeks | Ongoing (framework) | 4-8 weeks |
| Best suited for | EU/GDPR processing, French-regulated entities | UK processing, practical quick-start | US organisations, framework-based programs | Multinational, auditable, certification-seeking |
Methodology Selection Criteria
Decision Factors
| Factor | Weight | Considerations |
|---|---|---|
| Regulatory jurisdiction | High | Which supervisory authority will review the assessment? Use their preferred methodology. |
| Organisational maturity | Medium | Low maturity → ICO (simplest). Medium → CNIL. High → ISO 29134. |
| Processing complexity | Medium | Simple processing → ICO. Complex/high-risk → CNIL or ISO 29134. |
| International scope | High | Single jurisdiction → local DPA methodology. Multi-jurisdiction → ISO 29134. |
| Certification goals | Medium | Seeking ISO 27701 or Art. 42 certification → ISO 29134. |
| Resource availability | Medium | Limited resources → ICO. Dedicated privacy team → ISO 29134. |
| Existing framework | Low | Already using NIST CSF → add NIST PF. Already ISO 27001 → ISO 29134. |
Selection Decision Tree
- Is the processing subject to a specific supervisory authority that mandates or recommends a methodology?
- Yes → Use that authority's methodology (CNIL for France, ICO for UK).
- No → Continue to step 2.
- Does the organisation operate across multiple jurisdictions?
- Yes → ISO 29134 (internationally recognised).
- No → Continue to step 3.
- What is the organisation's privacy maturity level?
- Low (no formal privacy program) → ICO DPIA template (quickest to adopt).
- Medium (privacy program exists) → CNIL PIA tool (structured but accessible).
- High (mature privacy program, certification goals) → ISO 29134.
- Is the organisation US-based with an existing NIST CSF deployment?
- Yes → NIST PF as organisational framework, supplemented by CNIL or ICO for individual DPIAs.
- No → Use selection from steps 1-3.
Cross-Methodology Mapping
| CNIL Step | ICO Step | NIST PF Function | ISO 29134 Clause |
|---|---|---|---|
| Step 1: Context | Step 2: Describe processing | ID.IM (Inventory & Mapping) | Clause 6: Preparation |
| — | Step 1: Identify need | — | Clause 6.1: Determine necessity |
| — | Step 3: Consultation | CT.PO (Communication Policies) | Clause 6.4: Stakeholder engagement |
| Step 2: Fundamental Principles | Step 4: Necessity & proportionality | GV.PO (Governance Policies) | Clause 7.3: Privacy safeguard analysis |
| Step 3: Risks | Step 5: Identify & assess risks | ID.RA (Risk Assessment) | Clause 7.4: Risk analysis |
| Step 4: Validation | Step 6: Mitigation measures | CT.DM (Data Processing Management) | Clause 7.5: Risk treatment |
| — | Step 7: Sign off & record | GV.AT (Awareness & Training) | Clause 8: Follow-up |
Enforcement and Regulatory Guidance
- EDPB WP248rev.01: References both CNIL and ISO 29134 as acceptable DPIA methodologies. Does not mandate any specific methodology but requires that the chosen approach satisfies Art. 35(7)(a)-(d) minimum content.
- Art. 35(7) Minimum Content: Any methodology must include: (a) systematic description of processing and purposes; (b) necessity and proportionality assessment; (c) risk assessment; (d) measures to address risks. All four methodologies satisfy these requirements.
- CJEU C-175/20 (SS SIA 'Valsts ieņēmumu dienests'): Court confirmed that DPIA must assess risks from the perspective of data subjects, not the controller — aligns with ICO and ISO 29134 harm-based approaches.
- CNIL Deliberation 2018-327: Published criteria for acceptable DPIA methodologies — 10 criteria that any methodology must satisfy, regardless of which tool is used.
Conducting GDPR Data Protection Impact Assessment
Overview
A Data Protection Impact Assessment (DPIA) is a structured process mandated by Article 35 of the GDPR to identify, assess, and mitigate risks to the rights and freedoms of natural persons arising from data processing operations. The DPIA is not merely a compliance checkbox but a living risk management instrument that must be conducted before processing begins and updated throughout the processing lifecycle. This skill implements the methodology recommended by the European Data Protection Board in WP248rev.01 (Guidelines on Data Protection Impact Assessment) and incorporates enforceable requirements from Art. 35(1)-(11).
Mandatory DPIA Triggers — Art. 35(3)
A DPIA is required when processing is likely to result in a high risk to the rights and freedoms of natural persons, particularly in these cases:
Automatic Triggers Under Art. 35(3)
| Trigger | GDPR Reference | Description |
|---|---|---|
| Systematic and extensive profiling with significant effects | Art. 35(3)(a) | Automated processing including profiling that produces legal effects or similarly significant effects on the data subject |
| Large-scale special category or criminal data | Art. 35(3)(b) | Processing on a large scale of data under Art. 9(1) (health, biometric, genetic, racial, political, religious, sexual orientation, trade union) or Art. 10 (criminal convictions) |
| Systematic large-scale monitoring of public areas | Art. 35(3)(c) | CCTV surveillance, drone monitoring, Wi-Fi tracking, or other systematic monitoring of publicly accessible areas on a large scale |
EDPB WP248rev.01 Criteria — Two or More Triggers Require a DPIA
The EDPB established nine criteria for identifying high-risk processing. If a processing operation meets two or more of these criteria, a DPIA is presumptively required:
- Evaluation or scoring — Profiling, prediction, credit scoring, behavioural analysis
- Automated decision-making with legal or significant effect — Processing that determines access to services, contracts, or benefits
- Systematic monitoring — Observation, tracking, or surveillance of data subjects
- Sensitive data or data of a highly personal nature — Art. 9 special categories, financial data, communications metadata, location data
- Data processed on a large scale — Volume of data subjects, volume of data items, geographic scope, duration of processing
- Matching or combining datasets — Merging data from multiple sources beyond the data subject's reasonable expectation
- Data concerning vulnerable data subjects — Children, employees, mentally ill, asylum seekers, elderly, patients
- Innovative use or applying new technological solutions — Fingerprint and facial recognition combined with access control, IoT applications, AI-driven processing
- Processing that prevents data subjects from exercising a right or using a service — Including but not limited to screening processes that block access
DPIA Content Requirements — Art. 35(7)
Every DPIA must contain at minimum the following four elements:
(a) Systematic Description of Processing Operations and Purposes
- Nature of the processing: collection, recording, organisation, structuring, storage, adaptation, alteration, retrieval, consultation, use, disclosure, dissemination, combination, restriction, erasure, destruction
- Scope: categories of data subjects, categories of personal data, volume and frequency, geographic scope
- Context: relationship between controller and data subjects, how data was obtained, data subject expectations
- Purpose: specific purposes per Art. 5(1)(b), lawful basis under Art. 6(1), and if applicable Art. 9(2) condition
(b) Assessment of Necessity and Proportionality
- Lawful basis justification with specificity
- Purpose limitation analysis: is each data element necessary for the stated purpose?
- Data minimisation assessment: could the purpose be achieved with less data or anonymised data?
- Storage limitation: are retention periods justified and enforced?
- Data subject rights facilitation: how are Arts. 15-22 rights enabled?
- Safeguards for international transfers under Chapter V where applicable
(c) Assessment of Risks to Rights and Freedoms
For each identified risk, assess:
| Risk Dimension | Assessment Criteria |
|---|---|
| Likelihood | Remote (< 10%), Possible (10-50%), Likely (50-90%), Almost certain (> 90%) |
| Severity | Negligible (inconvenience), Limited (significant but recoverable), Significant (serious difficulty), Maximum (irreversible consequences) |
| Risk Level | Likelihood x Severity matrix producing Low, Medium, High, or Very High |
Types of harm to assess:
- Physical harm (discrimination leading to violence, denial of healthcare)
- Material harm (financial loss, identity theft, loss of employment)
- Non-material harm (reputational damage, emotional distress, loss of autonomy)
- Social harm (chilling effect on free speech, discrimination, exclusion)
- Loss of control over personal data (inability to exercise rights, unknown processing)
(d) Measures to Address Risks
- Technical measures: encryption at rest and in transit (AES-256, TLS 1.3), pseudonymisation per Art. 4(5), access controls (RBAC), automated deletion, audit logging
- Organisational measures: privacy policies, staff training, DPO oversight, incident response procedures, data processing agreements per Art. 28
- Contractual measures: data subject notification, consent mechanisms, processor obligations
- Residual risk assessment: risk level after implementation of mitigating measures
DPIA Process Methodology
Step 1: Screening and Scoping (Week 1)
- Complete the Privacy Threshold Analysis questionnaire to confirm DPIA obligation.
- Define the processing operation boundary: which systems, data flows, and organisational units are in scope.
- Identify the DPIA team: processing owner, DPO, IT security representative, legal counsel, and where appropriate a data subject representative per Art. 35(9).
- Gather existing documentation: system architecture diagrams, data flow maps, privacy notices, consent forms, processor agreements.
Step 2: Systematic Description (Week 2)
- Map the complete data lifecycle from collection to deletion.
- Document all data elements processed, with lawful basis for each.
- Identify all recipients and sub-processors.
- Document international transfers and applicable safeguard mechanisms.
- Record the technology stack and infrastructure involved.
Step 3: Necessity and Proportionality Assessment (Week 3)
- For each data element, document why it is necessary for the stated purpose.
- Assess whether less invasive alternatives could achieve the same purpose.
- Evaluate data minimisation compliance.
- Confirm retention periods are proportionate and technically enforced.
- Verify data subject rights can be exercised effectively.
Step 4: Risk Identification and Assessment (Week 3-4)
- Conduct threat modelling against the data flow map (sources of risk: internal actors, external attackers, processors, system failures).
- For each threat, assess likelihood and severity using the matrix in Art. 35(7)(c).
- Document risks in the risk register with unique identifiers.
- Calculate inherent risk levels before mitigation.
Step 5: Mitigation Measures (Week 4-5)
- For each High or Very High risk, identify specific technical and organisational measures.
- Document the expected risk reduction for each measure.
- Calculate residual risk levels after mitigation.
- If residual risk remains High or Very High, escalate to Art. 36 prior consultation with the supervisory authority.
Step 6: DPO Advice and Sign-Off (Week 5-6)
- Present the completed DPIA to the DPO for independent review per Art. 35(2).
- Document the DPO's advice and whether it was followed; if not, document reasons.
- Obtain processing owner sign-off and senior management approval.
- Record the DPIA in the central DPIA register with a scheduled review date.
Step 7: Ongoing Review
- Review the DPIA when there is a material change in risk: new data categories, new recipients, technology change, security incident, regulatory change.
- Conduct periodic reviews at minimum annually.
- Document all review outcomes and version the DPIA.
DPO Consultation — Art. 35(2)
The controller shall seek the advice of the Data Protection Officer where designated. The DPO must be involved from the screening phase. Per Art. 39(1)(c), the DPO has the task of providing advice regarding the DPIA and monitoring its performance. The controller must document the DPO's advice and the reasons for any departure from it.
Data Subject Views — Art. 35(9)
Where appropriate, the controller shall seek the views of data subjects or their representatives on the intended processing. This may take the form of:
- Surveys or focus groups with representative data subject samples
- Consultation with trade unions where employees are affected
- Public consultation for large-scale government processing
- Consultation with patient advocacy groups for health data processing
The controller must document whether data subject views were sought, and if not, the reasons why it was not appropriate.
Supervisory Authority Lists — Art. 35(4)-(5)
Each supervisory authority publishes a list of processing operations requiring a DPIA (Art. 35(4)) and may also publish a list of operations not requiring one (Art. 35(5)). Controllers must consult the applicable national list. Notable examples:
- CNIL (France): 14 processing types requiring DPIA, including employee scoring, health data warehouses, biometric access control
- ICO (UK): 10 processing types requiring DPIA, including invisible processing, tracking individuals online, automated decision-making
- BfDI (Germany): Processing of genetic or biometric data for identification, scoring by credit agencies, large-scale profiling
Common DPIA Deficiencies
- Generic risk descriptions: Risks described as "data breach" without specifying the attack vector, affected data categories, and specific harms
- Missing proportionality analysis: Jumping from description to risk without evaluating whether the processing is necessary and proportionate
- No residual risk calculation: Identifying measures but failing to assess whether they reduce risk to acceptable levels
- DPO advice not documented: DPO involvement limited to signature without recording substantive advice
- Static DPIA: Assessment completed once and never reviewed despite material changes to processing
- Missing Art. 35(9) justification: No documentation of whether data subject views were sought or why they were not
Enforcement Precedents
- Karolinska Institute (Swedish DPA, 2019): SEK 200,000 fine for processing genetic data without conducting a DPIA as required by Art. 35.
- Austrian Post (Austrian DPA, 2019): EUR 18 million fine related to profiling of political party affinities — DPIA inadequacy was a contributing factor.
- Clearview AI (CNIL, 2022): EUR 20 million fine for biometric processing without DPIA and without lawful basis.
- Real Madrid CF (AEPD, 2023): Sanctioned for implementing Wi-Fi tracking in stadium without conducting DPIA for large-scale public area monitoring under Art. 35(3)(c).
Integration Points
- Art. 36 Prior Consultation: When residual risk remains high after mitigation, the controller must consult the supervisory authority before processing begins.
- Art. 30 Records of Processing: DPIA findings should be cross-referenced with the RoPA to ensure consistency.
- Art. 25 Data Protection by Design: Mitigation measures identified in the DPIA feed directly into privacy-by-design implementation.
- Art. 33-34 Breach Notification: DPIA risk assessments inform breach severity analysis and notification decisions.
DPIA for Automated Decision-Making Systems
Purpose
Conduct a DPIA specifically tailored to automated decision-making (ADM) and profiling systems that produce legal or similarly significant effects on individuals, as required by GDPR Article 35(3)(a).
Prerequisites
- System documentation for the ADM/profiling system (algorithm design, training data, decision logic)
- Art. 35(3)(a) threshold confirmed: systematic and extensive evaluation of personal aspects based on automated processing, including profiling, on which decisions are based that produce legal effects or similarly significantly affect natural persons
- Data categories processed identified, including any Art. 9 special categories
- Existing Art. 22 safeguards documentation if available
Workflow
Step 1: System Description and Scope
Document the ADM system under assessment:
- System purpose: Business objective and processing purpose
- Decision types: What decisions the system makes or supports (credit scoring, insurance pricing, recruitment screening, fraud detection, content moderation)
- Processing scope: Data categories used as inputs, volume of data subjects affected, geographic scope
- Automation level: Fully automated (no human in loop), semi-automated (human in loop), or automated with human override capability
- Legal basis: Art. 6(1) basis for the processing, and if Art. 22 applies, the Art. 22(2) exception relied upon (contract, EU/Member State law, explicit consent)
Step 2: Necessity and Proportionality Assessment
Evaluate whether the automated processing is necessary and proportionate:
- Purpose limitation: Confirm processing purpose is specific, explicit, and legitimate per Art. 5(1)(b)
- Data minimization: Verify only data strictly necessary for the decision is processed per Art. 5(1)(c)
- Accuracy: Assess input data quality and its impact on decision accuracy per Art. 5(1)(d)
- Less intrusive alternatives: Document whether the decision could be made through less automated means
- Proportionality: Balance the organizational benefit against the impact on individuals' rights and freedoms
Step 3: Risk Identification
Identify risks specific to ADM systems:
- Discrimination risk: Assess whether the system could produce discriminatory outcomes based on Art. 9 characteristics (race, ethnicity, political opinions, religion, health, sexual orientation)
- Opacity risk: Evaluate whether the decision logic is explainable to affected individuals per Art. 13(2)(f) and Art. 14(2)(g)
- Accuracy risk: Assess false positive and false negative rates and their consequences
- Data quality risk: Evaluate whether training data reflects real-world populations without bias
- Feedback loop risk: Assess whether the system creates self-reinforcing biases
- Scope creep risk: Evaluate whether the system could be applied beyond its intended purpose
- Security risk: Assess adversarial manipulation vectors (data poisoning, model extraction)
Step 4: Art. 22 Compliance Assessment
If the system makes solely automated decisions with legal or similarly significant effects:
- Art. 22(1) prohibition check: Confirm whether the default prohibition applies
- Art. 22(2) exception: Document which exception permits the processing:
- (a) Necessary for entering into or performance of a contract
- (b) Authorized by EU or Member State law with suitable safeguards
- (c) Based on explicit consent
- Art. 22(3) safeguards: Verify implementation of mandatory safeguards:
- Right to obtain human intervention from the controller
- Right to express the data subject's point of view
- Right to contest the decision
- Art. 22(4) special categories: If Art. 9 data is processed, confirm explicit consent under Art. 9(2)(a) or substantial public interest under Art. 9(2)(g) with suitable safeguards
Step 5: Transparency and Explainability
Assess transparency obligations:
- Pre-decision transparency: Art. 13(2)(f) / Art. 14(2)(g) require providing meaningful information about the logic involved, significance, and envisaged consequences
- Logic explanation depth: Document what level of explanation is provided:
- General system functionality description
- Key factors/features influencing decisions
- Decision thresholds and their justification
- Individual-specific explanation of a particular decision
- Recital 71 guidance: Verify right to explanation is implemented as per Recital 71 (right to obtain an explanation of the decision reached after assessment)
Step 6: Risk Mitigation Measures
Implement measures to mitigate identified risks:
- Bias testing: Regular testing for discriminatory impact across protected characteristics
- Human oversight: Define when and how human reviewers intervene
- Accuracy monitoring: Ongoing performance monitoring with defined acceptable thresholds
- Audit trail: Complete logging of inputs, model version, and outputs for each decision
- Data subject rights: Clear process for individuals to exercise Art. 22(3) rights
- Regular review: Periodic re-assessment schedule (at minimum annually)
Step 7: Stakeholder Consultation
Consult required parties per Art. 35(9):
- DPO opinion: Obtain DPO assessment per Art. 35(2)
- Data subjects: Where appropriate, seek views of data subjects or their representatives per Art. 35(9)
- Technical experts: Algorithm developers, data scientists, domain experts
- Legal counsel: Assessment of applicable laws beyond GDPR (EU AI Act, sector-specific regulation)
Step 8: Documentation and Outcome
Document the DPIA outcome:
- Risk matrix: Residual risk rating after mitigation measures
- DPO recommendation: Documented DPO opinion on whether processing can proceed
- Prior consultation: If residual risk remains high, document decision to consult supervisory authority per Art. 36
- Review schedule: Next mandatory review date and trigger events for early review
Verification
- ADM system fully described with decision types, automation level, and data categories
- Art. 22(2) exception documented with Art. 22(3) safeguards implemented
- Discrimination risk assessed across Art. 9 protected characteristics
- Transparency obligations met per Art. 13(2)(f) / Art. 14(2)(g)
- Human intervention mechanism operational and tested
- Contestation process documented and accessible to data subjects
- DPO opinion obtained and documented
- Review schedule established with clear trigger events
DPIA Mitigation Planning
Overview
Article 35(7)(d) GDPR requires a DPIA to include "the measures envisaged to address the risks, including safeguards, security measures and mechanisms to ensure the protection of personal data and to demonstrate compliance with this Regulation." This skill provides a structured mitigation planning framework.
Mitigation Strategy Categories
Technical Measures
| Category | Examples | GDPR Reference |
|---|---|---|
| Encryption | At-rest, in-transit, end-to-end | Art. 32(1)(a) |
| Pseudonymisation | Tokenisation, hashing, key-coded | Art. 25(1), Art. 32(1)(a) |
| Access controls | RBAC, MFA, privileged access management | Art. 32(1)(b) |
| Data minimisation | Field-level reduction, aggregation, sampling | Art. 5(1)(c), Art. 25(1) |
| Anonymisation | k-anonymity, differential privacy, generalisation | Recital 26 |
| Monitoring | SIEM, DLP, anomaly detection | Art. 32(1)(d) |
Organisational Measures
| Category | Examples | GDPR Reference |
|---|---|---|
| Policies | Data protection policy, acceptable use | Art. 24(2) |
| Training | Privacy awareness, role-specific training | Art. 39(1)(b) |
| Contracts | DPAs, joint controller arrangements, NDAs | Art. 28, Art. 26 |
| Audits | Internal audits, processor audits, certification | Art. 28(3)(h) |
| Governance | DPO oversight, privacy committee, RACI | Art. 37-39 |
| Incident response | Breach procedures, notification protocols | Art. 33-34 |
Mitigation Plan Structure
For each identified risk:
- Risk reference -- Link to DPIA risk register entry
- Inherent risk level -- Before mitigation
- Proposed measures -- Specific technical and organisational controls
- Implementation owner -- Accountable person
- Implementation deadline -- Target completion date
- Verification method -- How effectiveness will be confirmed
- Residual risk level -- After planned mitigation
- Acceptance decision -- Within appetite / escalation required / prior consultation
Residual Risk Decision Framework
Residual Risk LOW → Accept; document; routine monitoring
Residual Risk MEDIUM → Accept with enhanced monitoring; annual review
Residual Risk HIGH → Escalate to senior management; consider additional measures
Residual Risk VERY HIGH → Art. 36 prior consultation required before processing
Implementation Tracking
Each mitigation measure progresses through:
- Proposed -- Identified but not yet approved
- Approved -- Budget and resources allocated
- In Progress -- Implementation underway
- Implemented -- Deployed and operational
- Verified -- Effectiveness confirmed through testing
DPIA Risk Scoring Methodology
Overview
Art. 35(7)(c) GDPR requires a DPIA to include "an assessment of the risks to the rights and freedoms of data subjects." This skill provides a quantifiable risk scoring framework that converts qualitative privacy risks into comparable, prioritised scores supporting mitigation decisions.
Risk Scoring Framework
Severity Scale (Impact on Data Subject Rights)
| Level | Score | Description | Examples |
|---|---|---|---|
| Negligible | 1 | Minor inconvenience, easily recoverable | Temporary inability to access non-essential service |
| Limited | 2 | Significant inconvenience, recoverable with effort | Targeted advertising based on inferred preferences |
| Significant | 3 | Serious consequences, difficult to recover from | Financial loss, discrimination, reputational harm |
| Maximum | 4 | Irreversible or very difficult to recover from | Identity theft, physical safety risk, loss of employment |
Likelihood Scale
| Level | Score | Description | Indicators |
|---|---|---|---|
| Negligible | 1 | Unlikely given current controls | Strong technical controls, limited access, encrypted at rest and in transit |
| Limited | 2 | Possible but requires specific conditions | Some access controls, partial encryption, known but unproven attack vectors |
| Significant | 3 | Probable given known threat landscape | Weak controls in specific areas, prior incidents in sector, active threat actors |
| Maximum | 4 | Near-certain or already occurring | No controls, known vulnerabilities, prior breach of similar system |
Risk Matrix
Severity → Negligible(1) Limited(2) Significant(3) Maximum(4)
Likelihood ↓
Maximum(4) 4(M) 8(H) 12(VH) 16(VH)
Significant(3) 3(L) 6(M) 9(H) 12(VH)
Limited(2) 2(L) 4(M) 6(M) 8(H)
Negligible(1) 1(L) 2(L) 3(L) 4(M)
Risk Levels: L=Low(1-3), M=Medium(4-6), H=High(7-9), VH=Very High(10-16)
Risk Categories (ENISA-Aligned)
Rights and Freedoms Impacts
- Loss of confidentiality -- Unauthorised disclosure of personal data
- Loss of integrity -- Unauthorised modification of personal data
- Loss of availability -- Inability to access or use personal data
- Loss of purpose limitation -- Data used beyond original purpose
- Discrimination -- Unfair treatment based on processing outcomes
- Identity theft/fraud -- Misuse of personal data for impersonation
- Financial loss -- Direct or indirect monetary harm
- Reputational damage -- Social standing or professional harm
- Physical harm -- Safety or health impacts
- Loss of autonomy -- Chilling effects on behaviour or free expression
Inherent vs Residual Risk
- Inherent risk: Risk level before applying any mitigation measures
- Residual risk: Risk level after applying planned mitigation measures
- Risk reduction: Difference between inherent and residual risk scores
- Acceptable risk threshold: Organisation-defined tolerance level
Art. 36 Prior Consultation Trigger
When residual risk remains High or Very High after all feasible mitigation measures, the controller must consult the supervisory authority under Art. 36(1) before commencing processing.
DPIA Stakeholder Consultation
Overview
Article 35(9) GDPR requires that controllers "shall, where appropriate, seek the views of data subjects or their representatives on the intended processing, without prejudice to the protection of commercial or public interests or the security of processing operations." This skill provides a structured approach to meeting this obligation.
Legal Requirements
Art. 35(9) -- Data Subject Consultation
The controller must seek data subject views "where appropriate." EDPB WP248rev.01 clarifies that failure to consult must be justified and documented.
Art. 35(2) -- DPO Involvement
"The controller shall seek the advice of the data protection officer, where designated, when carrying out a data protection impact assessment."
Art. 36(1) -- Prior Consultation Context
Where prior consultation with the supervisory authority is required, evidence of stakeholder consultation strengthens the submission.
Consultation Methods
| Method | Suitability | Scale | Cost |
|---|---|---|---|
| Focus groups | Deep qualitative insight | Small (8-12 participants) | Medium |
| Online surveys | Broad quantitative data | Large (hundreds+) | Low |
| Written consultation | Formal, documented responses | Medium | Low |
| Public meetings | Community engagement | Medium-Large | Medium |
| Representative body engagement | Sector-specific expertise | Varies | Low |
| Advisory panel | Ongoing input | Small (5-10 members) | Medium |
| User testing with privacy lens | Practical usability insight | Small-Medium | High |
Consultation Process
Step 1: Determine Appropriateness
- Assess whether consultation is feasible without compromising security
- Document any reasons for not consulting (Art. 35(9) exemption)
- Consider proportionality: scale of processing vs consultation effort
Step 2: Identify Stakeholders
- Direct data subjects (or representative sample)
- Data subject representative bodies (consumer associations, patient groups, trade unions)
- Data Protection Officer
- Processor representatives
- Relevant internal stakeholders (IT security, legal, business unit)
Step 3: Design Consultation
- Select appropriate method(s) from the table above
- Prepare accessible consultation materials explaining the processing
- Define the specific questions or areas for input
- Set timeline and response deadlines
Step 4: Execute and Document
- Conduct consultation activities
- Record all views received
- Analyse feedback for themes and concerns
- Document how feedback influenced the DPIA outcome
Step 5: Feedback Loop
- Inform consultees of the outcome and actions taken
- Explain where their views were incorporated or why they were not
- Maintain ongoing engagement channel for future DPIA reviews
Assessing Employee Surveillance Privacy
Overview
Workplace monitoring represents one of the most sensitive areas of data protection because of the inherent power imbalance between employer and employee. GDPR Art. 88 allows Member States to provide more specific rules for processing in the employment context, and WP29 Opinion 2/2017 on data processing at work provides detailed guidance on proportionality, transparency, and data minimisation for employee monitoring. This skill provides a DPIA methodology for all forms of workplace surveillance: email monitoring, internet usage tracking, CCTV, GPS vehicle and personnel tracking, keystroke and screen capture logging, telephone call monitoring, and wearable device monitoring.
Legal Framework
GDPR Art. 88 — Processing in the Employment Context
Art. 88(1) permits Member States to provide, by law or collective agreement, more specific rules for processing in the employment context, covering recruitment, performance of the employment contract, management, planning and organisation of work, equality and diversity, health and safety, protection of employer's property, and exercise of employment-related rights.
Art. 88(2) requires that such rules include suitable and specific measures to safeguard the data subject's human dignity, legitimate interests, and fundamental rights, with particular regard to transparency, intra-group transfers, and monitoring systems in the workplace.
WP29 Opinion 2/2017 on Data Processing at Work
Key principles established:
| Principle | Application to Workplace Monitoring |
|---|---|
| Necessity | Monitoring must be strictly necessary for the stated purpose; less intrusive alternatives must be evaluated first |
| Purpose limitation | Data collected for security monitoring cannot be repurposed for performance management without separate justification |
| Data minimisation | Monitoring should capture the minimum data necessary — metadata over content, aggregate over individual |
| Transparency | Employees must be clearly informed about the nature, scope, and purpose of monitoring before it begins |
| Proportionality | The intrusiveness of monitoring must be proportionate to the legitimate interest pursued; blanket monitoring is rarely proportionate |
| Consent limitations | Employee consent is generally not a valid lawful basis due to the power imbalance; legitimate interest (Art. 6(1)(f)) is more appropriate but requires rigorous balancing |
National Employment Data Protection Laws
| Country | Legislation | Key Provisions |
|---|---|---|
| Germany | BDSG Section 26 | Employee data processing only lawful when necessary for the employment relationship. Works council co-determination right (BetrVG Section 87(1)(6)) for technical monitoring equipment. |
| France | Labour Code L1121-1, L1222-4 | Proportionality and transparency requirements. CNIL Guidance on employee monitoring (2020). Prior works council consultation required. |
| Italy | Workers' Statute Art. 4 | Remote surveillance equipment prohibited unless agreed with trade unions or authorised by labour inspectorate. |
| Spain | LOPDGDD Art. 87-91 | Specific provisions for digital rights in the workplace, including right to digital disconnection. |
| Netherlands | UAVG Implementation Act | Works council consent required for employee monitoring systems per WOR Art. 27(1)(l). |
Monitoring Types and Proportionality Assessment
Email Monitoring
| Aspect | Assessment |
|---|---|
| Legitimate purposes | Regulatory compliance (FCA SYSC 10A, MiFID II), insider threat detection, intellectual property protection, harassment prevention |
| Proportionality levels | Metadata only (sender, recipient, timestamp) < Subject line + metadata < Full content scanning < Real-time content review |
| Data minimisation | Metadata monitoring is significantly less intrusive than content review; content should only be accessed with specific justification |
| Transparency | Employees must be informed of the scope of email monitoring in the privacy notice and employment contract |
| Private use | If employer permits private email use, monitoring must not extend to private communications; separation mechanisms should be implemented |
| Retention | Email monitoring data should have shorter retention than business email retention; monitoring metadata should be deleted within 30-90 days |
Internet Usage Monitoring
| Aspect | Assessment |
|---|---|
| Legitimate purposes | Network security (malware prevention), bandwidth management, acceptable use enforcement, regulatory compliance |
| Proportionality levels | Category-level monitoring < Domain-level logging < Full URL logging < Content inspection (DPI) |
| Data minimisation | Category-level classification is proportionate for most purposes; URL-level logging requires specific justification |
| Transparency | Employees must be informed of what is monitored and any automated blocking |
| Personal browsing | If personal internet use is permitted, monitoring should not capture personal browsing details; automatic exclusion of personal browsing categories where feasible |
CCTV Workplace Monitoring
| Aspect | Assessment |
|---|---|
| Legitimate purposes | Physical security, health and safety, theft prevention, regulatory compliance |
| Proportionality levels | External perimeter only < Reception and entry points < Common areas (excluding toilets, break rooms, changing areas) < Individual workstations |
| Prohibited areas | Toilets, changing rooms, break rooms, and prayer rooms must never be monitored by CCTV |
| Covert monitoring | Only permissible in exceptional circumstances (suspected criminal activity) for a limited period, with prior DPO approval and legal counsel sign-off |
| Retention | Maximum 30 days unless footage is required for a specific investigation; automated deletion |
| Audio recording | Generally disproportionate in the workplace; separate justification required |
GPS and Location Tracking
| Aspect | Assessment |
|---|---|
| Legitimate purposes | Fleet management, delivery route optimisation, lone worker safety, vehicle theft prevention |
| Proportionality | GPS tracking of company vehicles during working hours is generally proportionate; continuous tracking outside working hours is disproportionate |
| Data minimisation | Tracking frequency should be proportionate (periodic location updates vs continuous tracking); geofencing (alerts on boundary crossing) is less intrusive than continuous tracking |
| Personal vehicles | GPS tracking of employees' personal vehicles is highly intrusive and rarely proportionate |
| Working hours only | GPS tracking should automatically deactivate outside working hours or when vehicle is used for personal purposes (if permitted) |
Keystroke and Screen Capture Logging
| Aspect | Assessment |
|---|---|
| Legitimate purposes | Insider threat detection in high-security environments, fraud investigation |
| Proportionality | Keystroke logging is one of the most intrusive forms of monitoring — captures personal passwords, private messages, health searches, and intimate communications |
| Recommendation | Generally disproportionate for routine monitoring; may be justified only in specific high-risk roles (financial traders, classified information handlers) with strict access controls and short retention |
| DPO and WP29 position | WP29 Opinion 2/2017 states that technologies that monitor keystrokes or capture screenshots are very intrusive and should be considered a last resort |
DPIA Methodology for Employee Surveillance
Phase 1: Monitoring Justification (Week 1)
- Document the specific, legitimate purpose for each monitoring type.
- Assess whether the monitoring is necessary for the stated purpose or whether less intrusive alternatives exist.
- Conduct an Art. 6(1)(f) legitimate interest assessment (LIA) for each monitoring type:
- Controller's legitimate interest: what interest is being protected?
- Necessity: is monitoring necessary to achieve the interest?
- Balancing: do the employee's rights and freedoms override the legitimate interest?
- If using Art. 6(1)(c) legal obligation as lawful basis, identify the specific legal requirement (e.g., FCA SYSC 10A.1 for financial services communications recording).
Phase 2: Proportionality Balancing (Week 2)
For each monitoring type, apply the proportionality framework:
START: Proposed monitoring measure
│
├─ Is the monitoring measure necessary for the stated purpose?
│ ├─ NO → Monitoring is not justified. Remove from scope.
│ └─ YES → Continue.
│
├─ Could a less intrusive measure achieve the same purpose?
│ ├─ YES → Adopt the less intrusive measure instead.
│ └─ NO → Continue.
│
├─ Is the scope of monitoring proportionate?
│ ├─ Blanket monitoring of all employees?
│ │ └─ Generally disproportionate. Target monitoring to specific roles or risk areas.
│ ├─ Continuous monitoring during all working hours?
│ │ └─ Consider periodic or event-triggered monitoring instead.
│ └─ Monitoring extends to personal/private activity?
│ └─ Disproportionate. Implement exclusion mechanisms.
│
├─ Is the retention period proportionate?
│ ├─ Monitoring data retained indefinitely?
│ │ └─ Disproportionate. Set specific retention periods with automated deletion.
│ └─ Retention aligned with the monitoring purpose?
│ └─ Document justification for retention period.
│
└─ END: Document the proportionality assessment for each monitoring type.
Phase 3: Employee Consultation and Transparency (Week 3)
- Prepare employee privacy notice covering monitoring scope, purposes, and rights.
- Consult works council or employee representatives (mandatory in many jurisdictions):
- Germany: BetrVG Section 87(1)(6) co-determination right
- France: Prior consultation with CSE (Comite Social et Economique)
- Netherlands: Works council consent under WOR Art. 27(1)(l)
- Seek data subject views per Art. 35(9) through works council consultation.
- Update employment contracts and acceptable use policies.
Phase 4: Risk Assessment (Week 3-4)
Assess monitoring-specific risks:
| Risk | Description |
|---|---|
| Chilling effect | Employees modify legitimate behaviour due to awareness of monitoring, reducing creativity, communication, and wellbeing |
| Discriminatory application | Monitoring applied more intensively to certain employee groups based on role, location, or other characteristics |
| Personal data exposure | Monitoring captures personal communications, health information, or political/religious content |
| Repurposing | Data collected for security used for performance management or disciplinary purposes |
| Excessive access | Line managers or HR accessing individual monitoring data without justification |
| Psychological harm | Continuous monitoring causing stress, anxiety, and reduced job satisfaction |
Phase 5: Mitigation and Approval (Week 4-5)
- Implement technical and organisational measures to mitigate identified risks.
- Obtain DPO advice per Art. 35(2).
- Obtain works council agreement (where required).
- Senior management approval.
- Register DPIA and schedule review (6 months for employee monitoring, given sensitivity).
Enforcement Precedents
- CNIL vs Amazon France Logistique (2024): EUR 32 million fine for excessive employee surveillance in warehouses — scanner-based monitoring tracking idle time, stowing speed, and time between tasks was disproportionate to productivity management purposes.
- Italian Garante vs H&M (2020): EUR 35.3 million fine (Hamburg DPA) for systematic recording of private information about employees (health conditions, family problems, religious beliefs) during return-to-work interviews, creating detailed employee profiles.
- Greek DPA vs PwC Greece (2019): EUR 150,000 fine for unlawful employee email monitoring without adequate legal basis, transparency, or DPIA.
- Spanish AEPD vs Mercadona (2021): EUR 2.5 million fine for facial recognition of employees and customers — biometric surveillance without DPIA.
- Romanian DPA vs Raiffeisen Bank (2019): EUR 100,000 fine for GPS monitoring of employees' personal vehicles outside working hours.
- ECHR Barbulescu v Romania (Grand Chamber, 2017): Established that employee monitoring of communications must meet six criteria: prior notification, extent of monitoring, legitimate justification, less intrusive alternatives, consequences for employee, adequate safeguards.
Assessing Health Data Processing
Overview
Health data is classified as a special category under GDPR Art. 9(1), requiring both a lawful basis under Art. 6(1) and an Art. 9(2) exemption for processing. Processing of health data on a large scale triggers a mandatory DPIA under Art. 35(3)(b). This skill provides a DPIA methodology for health data processing scenarios: electronic health records, clinical trials, health research, telemedicine, health applications, medical devices, and genetic data analysis.
Legal Framework
Art. 9(1) — Definition of Health Data
GDPR Art. 4(15) defines data concerning health as "personal data related to the physical or mental health of a natural person, including the provision of health care services, which reveal information about his or her health status."
This includes:
- Medical records and diagnoses
- Prescription and medication data
- Laboratory results and imaging
- Mental health assessments
- Disability information
- Genetic data (Art. 4(13))
- Biometric data processed for health purposes
- Health insurance claims
- Fitness and wellness data that reveals health status
- Data inferred from non-health data that reveals health conditions
Art. 9(2) Exemptions for Health Data
| Exemption | Reference | Application |
|---|---|---|
| Explicit consent | Art. 9(2)(a) | Patient consent for specific health processing. Must be separate from treatment consent. |
| Employment obligations | Art. 9(2)(b) | Occupational health assessments required by employment law. |
| Vital interests | Art. 9(2)(c) | Emergency medical treatment where patient cannot consent. |
| Healthcare provision | Art. 9(2)(h) | Processing necessary for preventive or occupational medicine, medical diagnosis, provision of health or social care, or management of health systems. Must be processed by or under the responsibility of a professional subject to medical secrecy. |
| Public health | Art. 9(2)(i) | Processing necessary for public health purposes including protecting against serious cross-border threats, ensuring high standards of quality and safety of health care, medicinal products, or medical devices. |
| Scientific research | Art. 9(2)(j) | Processing necessary for scientific research purposes (including clinical trials) subject to Art. 89(1) safeguards: data minimisation, pseudonymisation where possible. |
EU Clinical Trials Regulation 536/2014
- Art. 28: Protection of subjects and informed consent.
- Art. 28(2): Informed consent must be separate from consent for data processing under GDPR.
- Art. 56: Transparency through the EU Clinical Trials Information System.
- Art. 61: Facility requirements including controlled access to clinical trial materials.
Health Data DPIA Methodology
Phase 1: Health Data Classification (Week 1)
- Inventory all health data elements processed.
- Classify each element:
- Direct health data (diagnoses, test results, prescriptions)
- Indirect health data (insurance claims, appointment records)
- Inferred health data (wellness app data, genetic risk scores)
- Genetic data (DNA sequences, genotype data, pharmacogenomic data)
- Identify the Art. 9(2) exemption for each processing activity.
- Verify that processing is under the responsibility of a health professional subject to professional secrecy (Art. 9(2)(h) and Art. 9(3)).
Phase 2: Lawful Basis and Exemption Analysis (Week 2)
For each health data processing activity, document:
| Requirement | Documentation |
|---|---|
| Art. 6(1) lawful basis | Which Art. 6(1)(a)-(f) basis applies |
| Art. 9(2) exemption | Which Art. 9(2)(a)-(j) exemption applies |
| Member State derogation | Whether national law provides additional conditions (Art. 9(4)) |
| Professional secrecy | Whether processing is under a professional subject to secrecy obligation (Art. 9(3)) |
| Data subject consent | If relying on Art. 9(2)(a), evidence of explicit consent separate from treatment consent |
Phase 3: Clinical Trial Specific Assessment (if applicable)
| Assessment Area | Requirements |
|---|---|
| Informed consent | CTR Art. 28 consent separate from GDPR consent; plain language; withdrawal right without impact on treatment |
| Pseudonymisation | Clinical trial data should be pseudonymised with key held by principal investigator; sponsor should not hold re-identification key |
| Data minimisation | Collect only data specified in the clinical trial protocol; no speculative data collection |
| International transfers | Multi-centre trials involve transfers; each transfer requires Chapter V mechanism |
| Retention | CTR Art. 58: sponsor retains trial master file for 25 years; reconcile with GDPR storage limitation |
| Secondary use | Re-use of clinical trial data for future research requires additional ethical approval and GDPR lawful basis |
Phase 4: HIPAA Crosswalk (for transatlantic operations)
| HIPAA Provision | GDPR Equivalent | Gap Assessment |
|---|---|---|
| HIPAA Privacy Rule (45 CFR 164.500-534) | GDPR Art. 5 principles + Art. 9 special categories | GDPR is more restrictive: broader definition of health data; explicit consent standard higher; no Treatment-Payment-Operations (TPO) exception |
| HIPAA Security Rule (45 CFR 164.302-318) | GDPR Art. 32 security of processing | Largely aligned: both require risk assessment, encryption, access controls, audit logging |
| HIPAA Breach Notification (45 CFR 164.400-414) | GDPR Art. 33-34 | GDPR stricter: 72-hour notification to SA; HIPAA allows 60 days |
| HIPAA Minimum Necessary | GDPR Art. 5(1)(c) data minimisation | Aligned in principle; GDPR applies more broadly |
| HIPAA Business Associate Agreement | GDPR Art. 28 DPA | Both require written agreements with processors; GDPR DPA is more prescriptive |
| HIPAA Patient Rights | GDPR Art. 15-22 | GDPR provides broader rights (portability, erasure, restriction, objection) |
| HIPAA De-identification (Safe Harbor / Expert) | GDPR Recital 26 anonymisation | Different standards: HIPAA Safe Harbor has 18 specific identifiers; GDPR requires irreversibility assessment |
Phase 5: Risk Assessment (Week 3-4)
| Risk | Description | Typical Level |
|---|---|---|
| HD-R1 | Health data breach exposing sensitive diagnoses, conditions, or genetic information | Very High |
| HD-R2 | Re-identification of pseudonymised health data through linkage with other datasets | High |
| HD-R3 | Discriminatory use of genetic data (insurance denial, employment discrimination) | High |
| HD-R4 | Unauthorised access by healthcare staff beyond clinical need-to-know | High |
| HD-R5 | Secondary use of clinical trial data without participant awareness | Medium |
| HD-R6 | Cross-border transfer of health data to jurisdictions with weaker protections | High |
| HD-R7 | Patient re-identification through aggregation of fitness/wellness app data | Medium |
Phase 6: Genetic Data Specifics
Genetic data has unique privacy characteristics:
- Perpetual sensitivity: Genetic information does not change over a lifetime and reveals information about biological relatives.
- Familial implications: An individual's genetic data reveals information about parents, siblings, and children who have not consented to processing.
- Predictive power: Genetic data can predict future health conditions, creating insurance and employment discrimination risks.
- Irreversibility: Once genetic data is disclosed, it cannot be "unlearned" or reset.
Additional requirements for genetic data:
- Art. 9(4) permits Member States to impose further conditions on genetic data processing.
- Germany: Gendiagnostikgesetz (GenDG) imposes strict limitations on genetic testing in employment and insurance contexts.
- France: Code de la sante publique Art. L1131-1 restricts genetic testing to medical and scientific purposes.
Enforcement Precedents
- Finnish DPA vs Pihlajalinna (2021): EUR 608,000 fine for health services company where employees accessed patient records without clinical justification — inadequate access controls and logging.
- Portuguese DPA vs Hospital do Barreiro (2018): EUR 400,000 fine for hospital where non-clinical staff had access to patient records through failure to implement role-based access controls.
- Dutch DPA vs OLVG Hospital (2021): EUR 440,000 fine for inadequate access controls on electronic patient records — insufficient two-factor authentication and logging of access.
- Swedish DPA vs Capio St. Goran (2020): SEK 30 million fine for hospital failing to conduct DPIA for patient records system and failing to implement adequate access controls.
- ICO vs Pharmacy2U (2015): GBP 130,000 fine for selling patient data to marketing companies without consent.
Assessing Marketing Analytics Privacy
Overview
Marketing analytics processing — including customer profiling, behavioural targeting, cross-device tracking, programmatic advertising, and conversion attribution — triggers multiple DPIA criteria under WP248rev.01: evaluation/scoring (C1), systematic monitoring (C3), matching or combining datasets (C6), and potentially innovative technology (C8). This skill provides a DPIA methodology for marketing analytics processing, integrating GDPR obligations with ePrivacy Directive requirements for cookie-based tracking and PECR compliance for UK-based operations.
Legal Framework
GDPR Provisions for Marketing Analytics
- Art. 5(1)(b) Purpose limitation: Marketing data collected for one purpose cannot be repurposed for incompatible marketing without further lawful basis.
- Art. 6(1)(a) Consent: Required for most marketing profiling; must be freely given, specific, informed, and unambiguous.
- Art. 6(1)(f) Legitimate interest: May apply to direct marketing to existing customers (Recital 47) but requires balancing test for profiling.
- Art. 21(2)-(3) Right to object: Data subjects have an absolute right to object to processing for direct marketing purposes, including profiling related to direct marketing.
- Art. 22 Automated decision-making: Profiling that produces legal or similarly significant effects requires Art. 22(2) exception and Art. 22(3) safeguards.
ePrivacy Directive (2002/58/EC) Art. 5(3)
Storing or accessing information on a user's terminal equipment (cookies, device fingerprinting, local storage) requires:
- Clear and comprehensive information about the purposes
- Consent of the user (interpreted per GDPR standard: freely given, specific, informed, unambiguous)
- Exception: strictly necessary cookies for the service explicitly requested by the user
PECR (UK Privacy and Electronic Communications Regulations 2003)
- Regulation 6: Cookie consent requirements (mirrors ePrivacy Art. 5(3)).
- Regulation 22: Unsolicited marketing communications require prior consent (opt-in) with exception for existing customer soft opt-in.
- ICO enforcement powers under Regulation 31.
Marketing Processing Types and Risk Assessment
Customer Profiling
| Aspect | Assessment |
|---|---|
| Description | Aggregating customer data to create profiles for segmentation and targeting |
| WP248 criteria | C1 (evaluation/scoring), C6 (matching datasets) |
| Lawful basis | Consent (Art. 6(1)(a)) for new prospects; legitimate interest (Art. 6(1)(f)) for existing customers with LIA |
| Key risks | Discriminatory profiling, unexpected inferences, purpose creep |
| Mitigation | Transparency about profiling logic; opt-out mechanism; regular profiling accuracy review |
Behavioural Targeting
| Aspect | Assessment |
|---|---|
| Description | Tracking online behaviour to serve targeted advertisements |
| WP248 criteria | C1 (scoring), C3 (systematic monitoring), C6 (matching), C8 (innovative tech) |
| Lawful basis | Consent required (ePrivacy Art. 5(3) for cookies + GDPR Art. 6(1)(a) for processing) |
| Key risks | Pervasive tracking, opaque adtech supply chain, data leakage to multiple parties |
| Mitigation | Consent management platform; vendor due diligence; real-time bidding data minimisation |
Cross-Device Tracking
| Aspect | Assessment |
|---|---|
| Description | Linking user activity across multiple devices (desktop, mobile, tablet, smart TV) |
| WP248 criteria | C1, C3, C6, C8 |
| Lawful basis | Consent required — cross-device tracking exceeds reasonable expectations |
| Key risks | Comprehensive behavioural profiling; re-identification of pseudonymous profiles; tracking beyond user awareness |
| Mitigation | Explicit consent for cross-device linking; device-level opt-out mechanisms; limited retention |
Conversion Attribution
| Aspect | Assessment |
|---|---|
| Description | Tracking user journey from ad impression to purchase to attribute marketing ROI |
| Lawful basis | Consent for cookie-based attribution; legitimate interest may apply for first-party server-side attribution |
| Key risks | Extended tracking windows; cross-site tracking; data sharing with attribution platforms |
DPIA Methodology for Marketing Analytics
Phase 1: Marketing Data Flow Mapping (Week 1)
- Inventory all marketing data sources (website analytics, CRM, email platform, social media, advertising platforms, DMP/CDP).
- Map data flows from collection to activation (profiling, targeting, measurement).
- Identify all third-party recipients (ad exchanges, demand-side platforms, data management platforms, social platforms).
- Document all cookies, pixels, and tracking technologies deployed.
- Identify cross-site and cross-device tracking mechanisms.
Phase 2: Lawful Basis Assessment (Week 2)
- For each marketing processing activity, determine the lawful basis:
- Cookie-based tracking: consent required (ePrivacy Art. 5(3))
- Profiling for targeting: consent (Art. 6(1)(a)) or legitimate interest with LIA (Art. 6(1)(f))
- Direct marketing emails: consent (PECR Reg. 22) or soft opt-in for existing customers
- Assess consent quality: is consent freely given, specific, informed, unambiguous, and withdrawable?
- For legitimate interest claims, conduct and document a legitimate interest assessment (LIA).
Phase 3: Risk Assessment (Week 3-4)
Assess marketing-specific risks:
| Risk | Description | Typical Level |
|---|---|---|
| MK-R1 | Opaque adtech supply chain — personal data shared with multiple parties without transparency | High |
| MK-R2 | Cross-site tracking building comprehensive browsing profiles beyond user expectation | High |
| MK-R3 | Discriminatory targeting — excluding or disadvantaging groups based on inferred characteristics | High |
| MK-R4 | Consent fatigue leading to uninformed consent | Medium |
| MK-R5 | Data leakage through real-time bidding bid requests | High |
| MK-R6 | Dark patterns in consent interfaces undermining genuine choice | High |
| MK-R7 | Children encountering targeted advertising | High |
| MK-R8 | Re-identification of pseudonymous marketing profiles | Medium |
Phase 4: Mitigation and Approval (Week 4-5)
- Implement technical measures: consent management platform, server-side analytics, data clean rooms, privacy sandbox APIs.
- Implement organisational measures: marketing data governance policy, vendor due diligence, data subject rights processes.
- DPO review and approval.
- Schedule review: annually or upon new marketing technology deployment.
Enforcement Precedents
- CNIL vs Google LLC (2022): EUR 150 million fine for making cookie rejection more difficult than acceptance on google.fr and youtube.com — dark pattern in consent interface.
- CNIL vs Amazon Europe (2020): EUR 35 million fine for placing advertising cookies without prior consent.
- CNIL vs Criteo (2023): EUR 40 million fine for behavioural advertising without valid consent, insufficient transparency about profiling, and failure to demonstrate consent had been obtained.
- Belgian DPA vs IAB Europe (2022): EUR 250,000 fine — Transparency and Consent Framework (TCF) consent string constitutes personal data; IAB Europe is a joint controller for TCF processing.
- Norwegian DPA vs Grindr (2021): NOK 65 million fine for sharing location and sexual orientation data with advertising technology partners without valid consent.
- AEPD vs CaixaBank (2020): EUR 6 million fine for commercial profiling without adequate consent management and insufficient transparency about profiling purposes.
- ICO vs TikTok (2023): GBP 12.7 million fine for processing children's data for targeted advertising without appropriate age verification and parental consent.
Assessing New Technology Privacy Impact
Overview
Article 35(1) of the GDPR explicitly references new technologies as a factor increasing the likelihood of high risk to data subjects' rights and freedoms. The EDPB in WP248rev.01 identifies innovative use of technology as one of nine criteria triggering a DPIA. This skill provides a structured PIA methodology for emerging technologies where the privacy implications are not yet fully understood, including Internet of Things (IoT), blockchain and distributed ledger technologies, augmented and virtual reality (AR/VR), quantum computing, digital twins, brain-computer interfaces, and ambient computing.
Technology-Specific Privacy Risk Profiles
Internet of Things (IoT)
| Risk Area | Description | Privacy Impact |
|---|---|---|
| Pervasive data collection | IoT devices continuously collect environmental and behavioural data, often without visible indicators | Data subjects may be unaware of the scope of data collection; transparency obligations under Art. 13-14 are difficult to fulfil on devices without screens |
| Data minimisation challenges | Sensors often collect more data than needed for the immediate purpose to enable future analytics | Violation of Art. 5(1)(c) data minimisation; purpose creep through accumulated data |
| Device-to-device communication | IoT ecosystems share data between devices without user awareness | Unexpected recipients; difficulty identifying all processors and sub-processors |
| Insecure by default | Many IoT devices ship with default credentials, unencrypted communications, and no update mechanism | Art. 25 data protection by design and Art. 32 security of processing obligations not met |
| Location and behavioural tracking | Connected devices reveal location patterns, daily routines, and behavioural habits | Systematic monitoring (WP248 C3); profiling risk (WP248 C1) |
| Cross-device correlation | Data from multiple IoT devices can be combined to create comprehensive behavioural profiles | Matching or combining datasets (WP248 C6); disproportionate surveillance |
EDPB Guidelines 02/2023 on IoT and Wearable Devices: Emphasized that IoT devices must implement data protection by design, provide clear privacy notices (adapted to device constraints), and enable genuine consent mechanisms.
Blockchain and Distributed Ledger Technology
| Risk Area | Description | Privacy Impact |
|---|---|---|
| Immutability vs right to erasure | Blockchain's append-only nature conflicts with Art. 17 right to erasure | Technical inability to delete personal data recorded on-chain |
| Transparency of transactions | Public blockchains expose transaction data to all participants | Personal data potentially accessible to unlimited recipients |
| Pseudonymity not anonymity | Blockchain addresses are pseudonymous but can be linked to real identities through transaction analysis | Re-identification risk; Art. 4(5) pseudonymisation does not equal anonymisation |
| Controller identification | Decentralised governance makes it difficult to identify the data controller | Art. 26 joint controller arrangements may be needed; accountability unclear |
| Cross-border by design | Distributed ledger nodes are typically located across multiple jurisdictions | Chapter V international transfer obligations triggered |
| Smart contract automation | Smart contracts execute automatically without human intervention | Art. 22 automated decision-making implications when smart contracts affect individuals |
CNIL Blockchain Guidance (2018): Recommended storing personal data off-chain with only hashes on-chain; using commitment schemes; designating participants who decide to use blockchain as controllers.
Augmented and Virtual Reality (AR/VR)
| Risk Area | Description | Privacy Impact |
|---|---|---|
| Biometric data collection | Eye tracking, facial expressions, body movements, voice patterns | Art. 9 special category data (biometric data for identification); Art. 35(3)(b) trigger |
| Spatial mapping | AR/VR devices scan and map physical environments including private spaces | Collection of data about third parties present in the environment without their consent |
| Behavioural profiling | Gaze tracking reveals interests, attention patterns, and cognitive state | Highly personal data; evaluation and scoring (WP248 C1) |
| Immersive manipulation | VR environments can influence behaviour through environmental design | Art. 5(1)(a) fairness; potential for subliminal manipulation |
| Persistent identity | Avatar and behavioural biometrics create persistent identifiable profiles | Long-term tracking across virtual environments |
| Child safety | Minors using VR platforms face enhanced risks | Vulnerable data subjects (WP248 C7); Art. 8 child consent requirements |
Quantum Computing
| Risk Area | Description | Privacy Impact |
|---|---|---|
| Cryptographic vulnerability | Quantum computers may break current encryption standards (RSA, ECC) | Art. 32 security measures based on current encryption become insufficient |
| Retroactive decryption | Encrypted data harvested today can be decrypted when quantum computers mature (harvest now, decrypt later) | Data currently protected may become exposed; long-term confidentiality compromised |
| Enhanced data analytics | Quantum machine learning can process data at scales impossible for classical computers | New forms of profiling and inference; privacy-preserving techniques may be defeated |
| Post-quantum migration | Transitioning to quantum-resistant cryptography requires significant infrastructure changes | Interim vulnerability period during migration |
ENISA Post-Quantum Cryptography Report (2024): Recommended organisations begin quantum risk assessment and plan migration to NIST-standardised post-quantum algorithms (ML-KEM, ML-DSA, SLH-DSA).
Digital Twins
| Risk Area | Description | Privacy Impact |
|---|---|---|
| Comprehensive data aggregation | Digital twins aggregate data from multiple sources to create a virtual replica | Matching or combining datasets (WP248 C6); comprehensive profiling |
| Predictive modelling of individuals | Digital twins of patients or employees predict future states and behaviours | Evaluation and scoring (WP248 C1); Art. 22 implications for predictions affecting individuals |
| Continuous synchronisation | Real-time data feeds maintain the digital twin's accuracy | Systematic monitoring (WP248 C3); proportionality concerns |
| Blurred anonymisation boundary | Even without direct identifiers, a sufficiently detailed digital twin may be re-identifiable | Pseudonymisation vs anonymisation assessment required |
PIA Methodology for Emerging Technologies
Phase 1: Technology Understanding (Week 1)
- Document the technology's data processing characteristics:
- What data is collected (categories, volume, frequency)
- How data flows through the technology stack
- Where data is stored and processed (geographic locations)
- Who has access to data at each stage
- How long data is retained
- What security measures protect data at each stage
- Identify the technology's novel risk characteristics that distinguish it from established technologies.
- Map the technology against the EDPB WP248rev.01 nine criteria.
- Consult published regulatory guidance specific to the technology (EDPB, CNIL, ICO, ENISA).
Phase 2: Stakeholder Impact Mapping (Week 2)
- Identify all affected data subjects:
- Direct users of the technology
- Third parties whose data may be incidentally collected
- Bystanders (particularly relevant for IoT, AR/VR, spatial computing)
- Identify vulnerable data subjects (children, employees, patients, elderly).
- Map potential impacts on each stakeholder group:
- Privacy impacts (loss of control, unexpected processing)
- Autonomy impacts (manipulation, reduced choice)
- Equality impacts (discriminatory outcomes, digital exclusion)
- Safety impacts (physical harm from decisions based on technology)
- Document data subject expectations versus actual processing.
Phase 3: Proportionality Assessment (Week 3)
For each data processing element of the technology:
| Assessment Question | Analysis Required |
|---|---|
| Is this processing necessary for the stated purpose? | Document why the technology cannot achieve its purpose without this data |
| Could the purpose be achieved with less data? | Evaluate data minimisation alternatives (aggregation, sampling, synthetic data) |
| Could the purpose be achieved with less identifying data? | Evaluate anonymisation, pseudonymisation, and differential privacy options |
| Could the purpose be achieved with a less invasive technology? | Compare the proposed technology against established alternatives |
| Are the benefits proportionate to the privacy intrusion? | Balancing test: public interest vs individual privacy impact |
| Have data subjects been consulted on the acceptability of the intrusion? | Art. 35(9) data subject views; user acceptance research |
Phase 4: Risk Assessment and Mitigation (Week 4-5)
- For each identified risk, assess likelihood and severity using the standard DPIA risk matrix.
- Identify technology-specific mitigation measures:
- Privacy-enhancing technologies (PETs): differential privacy, homomorphic encryption, secure multi-party computation, federated learning
- Data protection by design: privacy-preserving defaults, granular consent, purpose-bound processing
- Transparency mechanisms adapted to technology constraints (audio notices for screenless IoT, visual indicators for AR/VR)
- Assess residual risk after mitigation.
- If residual risk remains high, consider whether the technology should be deployed at all, not just how to mitigate risks.
Phase 5: Ongoing Monitoring Framework (Week 5-6)
- Define monitoring metrics specific to the technology:
- Data collection volumes (detect scope creep)
- Access patterns (detect unauthorised use)
- Accuracy and reliability of technology outputs
- User complaints and feedback
- Establish review triggers:
- Technology update or version change
- New use case or feature added
- Security vulnerability discovered
- Regulatory guidance issued for the technology
- Peer data breach involving similar technology
- Define a sunset plan: how will personal data be handled if the technology is decommissioned?
Privacy-by-Design Requirements for Emerging Technologies
| Principle | Implementation for Emerging Tech |
|---|---|
| Proactive not reactive | Conduct PIA before technology deployment, not after incidents |
| Privacy as default | Technology must ship with privacy-protective defaults; opt-in for additional data collection |
| Privacy embedded in design | Privacy requirements must be part of the technology specification, not bolt-on |
| Full functionality | Privacy protections should not degrade the technology's core functionality |
| End-to-end security | Data protection from collection through deletion, including inter-device communication |
| Visibility and transparency | Clear indicators when technology is collecting data; accessible privacy notices |
| Respect for user privacy | User-centric design; genuine choice and control over personal data |
Enforcement Precedents
- Spanish AEPD vs CaixaBank (2021): EUR 6 million fine for deploying new profiling technology (behavioural analytics for targeted financial products) without adequate DPIA for innovative technology use.
- ICO vs Clearview AI (2022): GBP 7.5 million fine for novel facial recognition technology scraping public images without lawful basis or DPIA.
- CNIL vs Amazon Europe (2020): EUR 35 million fine for deploying cookie tracking technology without valid consent on the Amazon.fr website.
- Italian Garante vs Foodinho/Glovo (2021): EUR 2.5 million fine for deploying algorithmic management technology (AI-driven shift allocation for riders) without DPIA.
- Belgian DPA vs IAB Europe (2022): EUR 250,000 fine for the Transparency and Consent Framework (TCF) used in programmatic advertising technology; found that TCF constituted processing of personal data without adequate controller designation.
Implementing NIST Privacy Framework IDENTIFY Function
Overview
The NIST Privacy Framework Version 1.0 (January 2020) provides a voluntary enterprise risk management tool structured around five core functions: IDENTIFY, GOVERN, CONTROL, COMMUNICATE, and PROTECT. The IDENTIFY function develops the organisational understanding to manage privacy risk arising from data processing. This skill covers the four IDENTIFY subcategories: Inventory and Mapping (ID.IM), Business Environment (ID.BE), Risk Assessment (ID.RA), and Data Processing Ecosystem Risk Management (ID.DE).
IDENTIFY Function Structure
ID.IM — Inventory and Mapping
Data processing inventory and mapping activities are understood by the organisation.
| Subcategory | Description | Implementation |
|---|---|---|
| ID.IM-P1 | Systems/products/services that process data are inventoried | Maintain a comprehensive register of all systems processing personal data, including SaaS tools, internal databases, and third-party integrations |
| ID.IM-P2 | Owners of systems/products/services are identified | Assign a data processing owner to each system with documented responsibility for privacy compliance |
| ID.IM-P3 | Categories of individuals whose data are processed are inventoried | Map all data subject categories (employees, customers, patients, vendors, website visitors) |
| ID.IM-P4 | Data actions of the systems/products/services are inventoried | Document data lifecycle actions: collection, retention, logging, generation, transformation, use, disclosure, sharing, transmission, disposal |
| ID.IM-P5 | Purposes for data actions are inventoried | Document specific purposes for each data action, aligned with GDPR Art. 5(1)(b) purpose limitation |
| ID.IM-P6 | Data elements within the data actions are inventoried | Catalogue all personal data elements processed per system and per purpose |
| ID.IM-P7 | Environmental factors affecting the data processing ecosystem are understood | Assess external factors: legal/regulatory requirements, industry standards, market expectations, organisational risk tolerance |
| ID.IM-P8 | Data processing is mapped | Create and maintain data flow maps showing how personal data moves through the organisation |
ID.BE — Business Environment
The organisation's mission, objectives, stakeholders, and activities are understood and prioritised.
| Subcategory | Description | Implementation |
|---|---|---|
| ID.BE-P1 | The organisation's role in the data processing ecosystem is identified and communicated | Determine controller/processor/joint controller status for each processing activity |
| ID.BE-P2 | Priorities for organisational mission, objectives, and activities are established and communicated | Align privacy programme with business objectives; ensure privacy is a design requirement |
| ID.BE-P3 | Systems/products/services that support organisational priorities are identified and key requirements communicated | Identify which processing activities are critical to business operations and prioritise privacy investment accordingly |
ID.RA — Risk Assessment
The organisation understands the privacy risks to individuals and how such processing creates privacy risks to the organisation.
| Subcategory | Description | Implementation |
|---|---|---|
| ID.RA-P1 | Contextual factors related to the systems/products/services and data actions are identified | Assess context: relationship with data subjects, data subject expectations, sensitivity, volume, collection method |
| ID.RA-P2 | Data analytic inputs and outputs are identified and evaluated | For data analytics and AI, identify inputs (training data, inference data) and outputs (predictions, scores, decisions) |
| ID.RA-P3 | Potential problems for individuals that arise from data processing are identified | Identify privacy harms: loss of autonomy, discrimination, economic loss, physical harm, reputational damage, loss of trust |
| ID.RA-P4 | Problematic data actions, the likelihood that they occur, and the resulting problems for individuals are identified and assessed | Assess likelihood and severity of each problematic data action — aligned with GDPR DPIA risk assessment |
| ID.RA-P5 | Risk responses are identified and prioritised | For each identified risk, determine response: mitigate, transfer, avoid, accept |
ID.DE — Data Processing Ecosystem Risk Management
The organisation identifies, assesses, and manages privacy risks associated with data processing by third parties.
| Subcategory | Description | Implementation |
|---|---|---|
| ID.DE-P1 | Data processing ecosystem risk management policies, processes, and procedures are identified, established, assessed, and managed | Vendor privacy management programme with due diligence, contractual requirements, and ongoing monitoring |
| ID.DE-P2 | Data processing ecosystem parties are identified, prioritised, and assessed | Map all processors, sub-processors, and third-party recipients with risk rating |
| ID.DE-P3 | Contracts with data processing ecosystem parties are established, maintained, and managed | Art. 28 DPAs with all processors; joint controller arrangements per Art. 26 |
| ID.DE-P4 | Interoperability frameworks or mechanisms are established for data processing ecosystem parties | Standard data formats, APIs, and privacy-preserving data exchange protocols |
| ID.DE-P5 | Data processing ecosystem parties are managed consistent with the organisation's risk strategy | Ongoing processor performance monitoring, audit rights exercise, sub-processor change management |
NIST PF to GDPR Mapping
| NIST PF Subcategory | GDPR Equivalent | Implementation Overlap |
|---|---|---|
| ID.IM-P1 through P8 | Art. 30 Records of Processing Activities | RoPA fulfils the core inventory and mapping requirements |
| ID.BE-P1 | Art. 26 Joint Controllers, Art. 28 Processors | Role determination aligns directly |
| ID.RA-P1 through P5 | Art. 35 DPIA | DPIA risk assessment methodology maps to NIST risk assessment |
| ID.DE-P1 through P5 | Art. 28 Processor Requirements | Processor governance maps to ecosystem risk management |
Implementation Methodology
Phase 1: Establish Inventory and Mapping (Months 1-3)
- Deploy a data discovery tool to identify personal data across all systems.
- Conduct departmental interviews to identify informal data processing.
- Create the system-level processing inventory (ID.IM-P1, P2).
- Map data subject categories (ID.IM-P3) and data elements (ID.IM-P6).
- Document data actions per system (ID.IM-P4) and purposes (ID.IM-P5).
- Create data flow maps (ID.IM-P8).
- Cross-reference with existing Art. 30 RoPA.
Phase 2: Business Environment Analysis (Month 2-3)
- Determine controller/processor/joint controller roles (ID.BE-P1).
- Align privacy programme priorities with business strategy (ID.BE-P2).
- Identify critical processing activities (ID.BE-P3).
Phase 3: Risk Assessment (Months 3-5)
- For each processing activity, assess contextual factors (ID.RA-P1).
- Identify problematic data actions using the NIST problematic data action taxonomy:
- Appropriation: use of data in ways beyond data subject expectations
- Distortion: use of inaccurate data or misleading inferences
- Induced disclosure: pressure to reveal more data than intended
- Insecurity: inadequate protection of data
- Re-identification: linking anonymised data back to individuals
- Stigmatisation: association with disfavoured groups
- Surveillance: excessive monitoring
- Unanticipated revelation: discovery of information beyond original data
- Unwarranted restriction: limiting access to services based on data
- Assess likelihood and severity for each problematic data action (ID.RA-P4).
- Determine risk responses (ID.RA-P5).
Phase 4: Ecosystem Risk Management (Months 4-6)
- Inventory all third-party data processors and recipients (ID.DE-P2).
- Conduct vendor privacy due diligence assessments.
- Verify DPA compliance for all processors (ID.DE-P3).
- Establish ongoing monitoring programme (ID.DE-P5).
Enforcement Context
While the NIST Privacy Framework is voluntary and not directly enforceable, organisations demonstrating NIST PF implementation show a mature privacy programme that can support GDPR accountability under Art. 5(2) and Art. 24. US state privacy laws (California CPRA, Virginia VCDPA, Colorado CPA) increasingly reference the NIST PF as a compliance resource.
Managing PIA Review and Update Cadence
Overview
Art. 35(11) requires controllers to carry out a review to assess whether processing is performed in accordance with the DPIA, at least when there is a change in the risk represented by processing operations. A DPIA is not a one-time compliance exercise but a living document that must be maintained throughout the processing lifecycle. This skill establishes the governance framework for DPIA reviews, including trigger event identification, scheduled review cadence, version control, stakeholder sign-off, and DPIA register management.
Art. 35(11) Review Obligation
"Where necessary, the controller shall carry out a review to assess if processing is performed in accordance with the data protection impact assessment at least when there is a change in the risk represented by processing operations."
This means:
- Reviews are mandatory when risk changes
- The controller should proactively monitor for risk changes
- Regular periodic reviews are a best practice even without identified changes
- The DPIA must be updated to reflect the current state of processing
Review Trigger Categories
Category 1: Regulatory Triggers
| Trigger | Assessment Action |
|---|---|
| New GDPR guidance from EDPB or national supervisory authority | Assess whether guidance changes the interpretation of lawful basis, proportionality, or risk level for the processing |
| Adequacy decision adopted, renewed, or invalidated | Reassess international transfers documented in the DPIA |
| National supervisory authority DPIA list updated (Art. 35(4)) | Check if processing now appears on the mandatory DPIA list |
| New Member State employment, health, or sector-specific law | Assess impact on lawful basis and proportionality for affected processing |
| Enforcement decision against similar processing by same or different SA | Assess whether the enforcement rationale applies to the organisation's processing |
| AI Act obligations becoming applicable | For AI-related DPIAs, assess conformity requirements |
Category 2: Processing Change Triggers
| Trigger | Assessment Action |
|---|---|
| New personal data category added | Reassess data minimisation, proportionality, and risk levels |
| New category of data subjects | Reassess impact on new data subject group; check for vulnerable subjects |
| New recipient or processor | Update data flow map; reassess sub-processor chain; verify DPA |
| Change in lawful basis | Full reassessment of lawful basis justification and proportionality |
| Change in retention period | Reassess storage limitation compliance |
| New purpose identified for existing data | Art. 6(4) compatibility assessment; update DPIA purpose description |
| Processing volume significantly increased | Reassess whether processing now qualifies as "large scale" |
| Geographic scope expanded | Assess new jurisdictions for data protection requirements and international transfers |
Category 3: Technology Change Triggers
| Trigger | Assessment Action |
|---|---|
| New system or platform deployment | Full technical risk reassessment; update system description and data flow maps |
| System upgrade or migration | Assess whether changes affect security measures, data flows, or processing scope |
| New algorithm or AI model deployed | Reassess algorithmic fairness, Art. 22 applicability, and AI Act classification |
| New tracking or monitoring technology | Reassess proportionality and ePrivacy compliance |
| Security vulnerability discovered | Immediate risk reassessment; assess whether existing mitigations remain effective |
| End-of-life for system component | Assess impact of reduced vendor support on security posture |
Category 4: Incident Triggers
| Trigger | Assessment Action |
|---|---|
| Personal data breach involving the processing activity | Mandatory DPIA review; reassess risk levels and mitigation effectiveness |
| Near-miss security incident | Assess whether existing mitigations prevented the near-miss or were bypassed |
| Data subject complaint about the processing | Investigate whether complaint reveals processing not documented in DPIA |
| Supervisory authority enquiry about the processing | Review DPIA for completeness and accuracy before responding |
| Internal audit finding | Address audit findings and update DPIA accordingly |
Category 5: Organisational Change Triggers
| Trigger | Assessment Action |
|---|---|
| Merger, acquisition, or divestiture | Reassess controller identity, data sharing arrangements, and international transfers |
| DPO change | New DPO to review and approve existing DPIAs |
| Significant staff restructuring | Verify processing owners, access controls, and organisational measures remain appropriate |
| New business unit using the processing | Assess whether new use case was documented in the DPIA |
Scheduled Review Cadence
| Processing Risk Level | Recommended Review Interval | Justification |
|---|---|---|
| Very High (prior consultation was required) | Every 6 months | Processing that required Art. 36 prior consultation needs frequent monitoring |
| High (biometric, large-scale special category, AI decision-making) | Every 6-12 months | High-risk processing warrants at least annual review |
| Medium (standard employee monitoring, marketing profiling) | Every 12 months | Annual review is the standard best practice |
| Low (routine processing with low residual risk) | Every 24 months | Extended review interval acceptable for stable, low-risk processing |
Review Process
Step 1: Trigger Detection and Triage
- DPIA register manager monitors for trigger events continuously.
- When a trigger is detected, it is logged in the DPIA register against the affected DPIA(s).
- DPO triages the trigger to determine whether a full or targeted review is needed:
- Full review: Re-examine all DPIA sections (new system, major regulatory change, data breach)
- Targeted review: Update only the affected section (new recipient, retention change, minor system update)
Step 2: Review Execution
- Assemble the review team: processing owner, DPO, IT security (for technology changes), legal (for regulatory changes).
- For full reviews: re-execute the complete DPIA methodology from screening through risk assessment.
- For targeted reviews: update the affected section and reassess overall risk level.
- Document all changes from the previous version.
- Update the risk register with any new or modified risks.
Step 3: Stakeholder Sign-Off
| Stakeholder | Role in Review | Sign-Off Required? |
|---|---|---|
| Processing Owner | Validates accuracy of processing description and confirms mitigations are implemented | Yes |
| DPO | Provides independent advice on risk assessment and proportionality per Art. 35(2) | Yes |
| IT Security / CISO | Validates technical measures and security risk assessment | Yes (for technology-related reviews) |
| Legal Counsel | Validates lawful basis and regulatory compliance | Yes (for regulatory trigger reviews) |
| Senior Management | Approves residual risk acceptance | Yes (for full reviews or risk level changes) |
| Works Council | Consulted on employee monitoring changes (jurisdiction-dependent) | Yes (for employee monitoring DPIAs) |
Step 4: Version Control and Registration
- Create a new version of the DPIA document.
- Use version numbering: major changes increment the first digit (v2.0); minor changes increment the second digit (v1.1).
- Include a change log documenting:
- Date of change
- Trigger event that prompted the review
- Sections modified
- Summary of changes
- Reviewer and approver names
- Archive the previous version (do not delete; maintain audit trail).
- Update the DPIA register with the new version, review date, and next review date.
DPIA Register Management
The DPIA register is the central index of all DPIAs conducted by the organisation. It must contain:
| Field | Description |
|---|---|
| DPIA Reference | Unique identifier (DPIA-[ORG]-[YEAR]-[SEQ]) |
| Processing Activity Name | Name of the processing operation |
| Version | Current version number |
| Status | Draft, Under Review, Approved, Expired, Superseded |
| Date Created | Date of initial DPIA |
| Last Review Date | Date of most recent review |
| Next Review Date | Scheduled date for next review |
| Processing Owner | Name and role of the processing owner |
| DPO Reviewer | Name of the DPO who reviewed |
| Risk Level (Current) | Current overall risk level after mitigation |
| Prior Consultation Required | Whether Art. 36 was triggered |
| Related Processing Activities | Cross-references to related DPIAs |
| Change Log | Summary of version history |
Common Review Process Deficiencies
- No trigger monitoring: Organisation relies solely on scheduled reviews without monitoring for change events.
- Stale DPIAs: Reviews conducted on paper without verifying actual processing matches the DPIA description.
- No version control: Edits made to the original document without maintaining previous versions.
- Incomplete stakeholder engagement: DPO reviews without consulting the processing owner or IT security.
- No change log: Updates made without documenting what changed and why.
- Expired review dates: DPIA register shows overdue reviews without remediation.
Enforcement Context
Supervisory authorities have increasingly scrutinised DPIA currency:
- Swedish DPA vs Karolinska (2019): Noted that DPIA had not been conducted at all, let alone reviewed.
- CNIL vs Clearview AI (2022): The absence of any DPIA was an aggravating factor in the fine determination.
- Austrian DPA vs Austrian Post (2019): Inadequate ongoing review of profiling DPIA was a contributing factor.
PIA Threshold Screening
Overview
Article 35(1) GDPR requires a DPIA when processing is "likely to result in a high risk to the rights and freedoms of natural persons." This skill provides a structured screening methodology to make that determination before committing to a full DPIA.
Screening Criteria
EDPB Nine-Criteria Test (WP248rev.01)
Processing that meets two or more of the following criteria generally requires a DPIA:
| # | Criterion | GDPR Reference |
|---|---|---|
| 1 | Evaluation or scoring | Art. 35(3)(a) |
| 2 | Automated decision-making with legal/similar effect | Art. 35(3)(a) |
| 3 | Systematic monitoring | Art. 35(3)(c) |
| 4 | Sensitive data or data of highly personal nature | Art. 9, Art. 10 |
| 5 | Data processed on a large scale | Recital 91 |
| 6 | Matching or combining datasets | WP248 |
| 7 | Data concerning vulnerable data subjects | WP248 |
| 8 | Innovative use or applying new technological solutions | WP248 |
| 9 | Processing that prevents data subjects from exercising a right | Art. 22, Art. 35(3)(b) |
Mandatory DPIA Triggers (Art. 35(3))
A DPIA is always required for:
- (a) Systematic and extensive evaluation of personal aspects based on automated processing, including profiling, producing legal effects or similarly significant effects
- (b) Large-scale processing of special categories of data (Art. 9(1)) or criminal conviction data (Art. 10)
- (c) Systematic monitoring of a publicly accessible area on a large scale
National Supervisory Authority Blacklists (Art. 35(4))
Each EU/EEA supervisory authority publishes a list of processing operations requiring a DPIA. The screening must check the relevant national blacklist based on the controller establishment.
Screening Process
- Identify processing activity -- Describe the proposed or changed processing
- Check mandatory triggers -- Evaluate against Art. 35(3)(a)-(c)
- Check national blacklist -- Consult the relevant SA published list
- Apply nine-criteria test -- Score each WP248 criterion as met/not met
- Document rationale -- Record the screening decision with justification
- Escalate or close -- Route to full DPIA or document exemption
Decision Logic
IF any Art. 35(3) mandatory trigger is met → DPIA REQUIRED
ELSE IF processing appears on national SA blacklist → DPIA REQUIRED
ELSE IF 2+ WP248 criteria are met → DPIA REQUIRED
ELSE IF 1 WP248 criterion is met → DPIA RECOMMENDED (risk-based decision)
ELSE → DPIA NOT REQUIRED (document exemption)
Integration Points
- RoPA linkage: Screening results attach to the corresponding RoPA entry
- Change management: New processing or material changes trigger re-screening
- DPIA register: Positive screening decisions feed the DPIA pipeline
Conducting Prior Consultation with Supervisory Authority
Overview
Article 36(1) requires the controller to consult the supervisory authority prior to processing where a DPIA under Art. 35 indicates that the processing would result in a high risk in the absence of measures taken by the controller to mitigate the risk. Prior consultation is the final safety net when a DPIA reveals risks that cannot be adequately mitigated through technical, organisational, or contractual measures. The supervisory authority has up to 8 weeks (extendable by 6 weeks) to provide written advice. Processing must not commence until the supervisory authority's response is received and addressed.
When Prior Consultation Is Required
Art. 36(1) Trigger
Prior consultation is mandatory when:
- A DPIA has been conducted under Art. 35
- The DPIA identifies that processing would result in high risk
- The controller cannot sufficiently mitigate the risk through available measures
- Residual risk remains High or Very High despite all reasonable mitigations
Residual Risk Assessment
| Residual Risk Level | Action |
|---|---|
| Low | No prior consultation needed. Processing may proceed. |
| Medium | No prior consultation required, but DPO should document the acceptance rationale. |
| High | Prior consultation recommended. Processing should not proceed without SA review. |
| Very High | Prior consultation mandatory under Art. 36(1). Processing must not proceed. |
Practical Examples Triggering Prior Consultation
- Large-scale biometric identification in publicly accessible areas where encryption with exporter-held key is not feasible
- AI system making automated decisions about access to essential services (credit, insurance, healthcare) where residual discrimination risk cannot be fully eliminated
- Processing of genetic data for a novel purpose where no existing Art. 9(2) exemption clearly applies
- Employee monitoring system where proportionality concerns remain despite all mitigations
- International transfer to a country with Very High government access risk where the processing requires clear-text data access
Art. 36(3) Required Documentation
The controller must provide the supervisory authority with the following information:
| Document | Content | Art. 36(3) Reference |
|---|---|---|
| DPIA report | Complete DPIA per Art. 35(7) including systematic description, necessity/proportionality, risk assessment, and mitigation measures | Art. 36(3)(a) — implicit |
| Controller and processor responsibilities | Clear documentation of which entity is responsible for which processing operations and security measures | Art. 36(3)(a) |
| Measures and safeguards | All technical, organisational, and contractual measures implemented to protect data subjects | Art. 36(3)(b) |
| DPO contact details | Contact details of the Data Protection Officer | Art. 36(3)(c) |
| Additional information requested | Any further information the supervisory authority requests during the consultation | Art. 36(3)(d) |
Supplementary Documentation (Best Practice)
| Document | Purpose |
|---|---|
| Executive summary | 1-2 page summary of the processing, identified risks, and reasons for prior consultation |
| Data flow diagram | Visual representation of personal data flows through the processing |
| Risk register | Detailed risk register with inherent and residual risk levels |
| Mitigation measures schedule | Implementation status and timeline for all mitigations |
| DPO advice letter | Written DPO advice per Art. 35(2) and whether it was followed |
| Legitimate interest assessment | If processing relies on Art. 6(1)(f), the documented LIA |
| Art. 35(9) consultation evidence | Documentation of data subject views sought |
Consultation Process
Step 1: Preparation (Weeks 1-2)
- DPO confirms that prior consultation is required based on DPIA residual risk assessment.
- Processing owner compiles the Art. 36(3) documentation package.
- DPO reviews the package for completeness and accuracy.
- Legal counsel reviews for legal accuracy and strategic considerations.
- Senior management is briefed on the prior consultation and its implications (processing pause).
Step 2: Submission (Week 3)
- Identify the competent supervisory authority:
- Lead SA under one-stop-shop mechanism (Art. 56) for cross-border processing
- Local SA for purely national processing
- Submit the consultation package through the SA's designated channel (online portal, registered mail, or both per SA requirements).
- Confirm receipt and obtain a case reference number.
- Document the submission date — this starts the 8-week clock.
Step 3: Supervisory Authority Review Period (Weeks 3-11)
Art. 36(2) timeline:
- Standard period: Up to 8 weeks from receipt of the complete consultation request
- Extension: The SA may extend the period by up to 6 weeks for complex cases, with notification to the controller within the first month
- Total maximum: 14 weeks (8 + 6)
During this period:
- The SA may request additional information under Art. 36(3)(d). The 8-week clock pauses until the information is provided.
- The DPO serves as the point of contact for SA communications.
- Processing must not commence during the consultation period.
- The controller should continue implementing mitigation measures in preparation.
Step 4: Outcome Management
| SA Response | Controller Action |
|---|---|
| Written advice with no concerns | Document the SA's response. Proceed with processing. Update DPIA with SA clearance. |
| Written advice with recommendations | Implement the SA's recommendations. Update DPIA. Proceed with processing once recommendations are addressed. |
| Written advice indicating GDPR infringement | Do not proceed with processing as described. Review processing design to address SA concerns. Consider re-submission after modifications. |
| No response within the statutory period | SA silence does not constitute approval. The controller bears responsibility for demonstrating compliance. Proceed with caution, documenting the SA's non-response. |
| Request for more information | Provide the requested information promptly. The clock pauses. |
Step 5: Post-Consultation
- Document the entire consultation process and outcome in the DPIA register.
- Update the DPIA with any SA recommendations or requirements.
- If processing proceeds, implement all SA-recommended measures.
- Schedule an accelerated DPIA review (6 months post-processing commencement).
- Maintain the consultation documentation for supervisory authority inspection.
National Supervisory Authority Consultation Procedures
| SA | Submission Method | Specific Requirements |
|---|---|---|
| ICO (UK) | Online consultation request form | Include DPIA, data flow diagram, DPO contact, description of why risk cannot be mitigated |
| CNIL (France) | Online portal (teleservice.cnil.fr) | DPIA in CNIL format; French language required |
| BfDI (Germany) | Written submission to the competent state or federal DPA | German language; DPIA must reference BDSG provisions |
| Garante (Italy) | PEC (certified email) | Italian language; include DPIA, processor agreements, security measures documentation |
| AEPD (Spain) | Sede Electronica portal | Spanish language; include DPIA and all supporting documentation |
| DPC Ireland | Online submission through DPC website | English language; include DPIA and all Art. 36(3) documentation |
Common Prior Consultation Mistakes
- Consulting too early: Submitting before completing the DPIA and implementing all feasible mitigations. The SA expects to see that all reasonable measures have been taken.
- Consulting too late: Beginning processing before the SA responds. Processing must not start during the consultation period.
- Incomplete documentation: Submitting without all Art. 36(3) required documents, causing delays for information requests.
- No DPO involvement: Submitting without documented DPO advice per Art. 35(2).
- Treating SA silence as approval: SA non-response does not constitute approval or endorsement.
- Failing to implement SA recommendations: Proceeding with processing without addressing SA feedback.
Enforcement Context
- Art. 36(1) non-compliance (failure to consult when required) can result in administrative fines up to EUR 10 million or 2% of annual worldwide turnover under Art. 83(4)(a).
- Swedish DPA vs Karolinska (2019): Fine included failure to conduct DPIA and consequent failure to identify need for prior consultation.
- Norwegian DPA vs Municipality (2020): Enforcement action for proceeding with high-risk processing without prior consultation.
Conducting Privacy Threshold Analysis
Overview
A Privacy Threshold Analysis (PTA) is a lightweight screening tool used to determine whether a processing activity requires a full DPIA under Art. 35. The PTA functions as a triage mechanism: it applies the WP248rev.01 nine criteria, Art. 35(3) mandatory triggers, and national supervisory authority DPIA lists to quickly classify processing activities into three categories: DPIA required, DPIA recommended, or DPIA not required. Every new processing activity, system change, or procurement of data-processing services should pass through the PTA before implementation.
PTA Quick-Screen Questionnaire
Section A: Art. 35(3) Mandatory Triggers
| Question | Yes/No | If Yes |
|---|---|---|
| A1. Does the processing involve systematic and extensive evaluation of personal aspects based on automated processing (including profiling) on which decisions are based that produce legal effects or similarly significantly affect individuals? | DPIA mandatory — Art. 35(3)(a) | |
| A2. Does the processing involve large-scale processing of special categories of data (Art. 9(1): health, biometric, genetic, racial/ethnic, political, religious, trade union, sexual orientation) or criminal conviction data (Art. 10)? | DPIA mandatory — Art. 35(3)(b) | |
| A3. Does the processing involve systematic monitoring of a publicly accessible area on a large scale (e.g., CCTV in public spaces, Wi-Fi tracking)? | DPIA mandatory — Art. 35(3)(c) | |
| A4. Does the processing appear on the national supervisory authority's DPIA required list (Art. 35(4))? | DPIA mandatory |
If any question in Section A is answered Yes: DPIA is mandatory. Stop. Proceed to full DPIA.
Section B: EDPB WP248rev.01 Nine Criteria
| Question | Criterion | Yes/No |
|---|---|---|
| B1. Does the processing involve evaluation or scoring of individuals (profiling, prediction, credit scoring, behavioural analysis)? | C1 | |
| B2. Does the processing involve automated decision-making that produces legal effects or similarly significant effects on individuals? | C2 | |
| B3. Does the processing involve systematic monitoring of individuals (observation, tracking, surveillance)? | C3 | |
| B4. Does the processing involve sensitive data (Art. 9 special categories) or highly personal data (financial, location, communications)? | C4 | |
| B5. Is the processing carried out on a large scale (number of subjects, volume, geographic scope, duration)? | C5 | |
| B6. Does the processing involve matching or combining datasets from different sources beyond data subject expectations? | C6 | |
| B7. Does the processing involve data concerning vulnerable individuals (children, employees, patients, elderly, disabled, asylum seekers)? | C7 | |
| B8. Does the processing involve innovative use of technology or application of existing technology in a new way? | C8 | |
| B9. Does the processing prevent individuals from exercising a right or using a service or contract? | C9 |
Scoring:
- 2 or more criteria met: DPIA is strongly recommended (presumptively required per WP248rev.01)
- 1 criterion met: DPIA is recommended. Consult DPO for final determination.
- 0 criteria met: DPIA is not required based on EDPB criteria.
Section C: Additional Risk Factors
| Question | Yes/No | Impact |
|---|---|---|
| C1. Does the processing involve international transfers to countries without an adequacy decision? | Additional risk factor — TIA required alongside DPIA if applicable | |
| C2. Is this a new processing activity not previously assessed? | New processing should be assessed more conservatively | |
| C3. Has similar processing been subject to enforcement action by any supervisory authority? | Elevated scrutiny — DPIA recommended regardless of criteria count | |
| C4. Does the processing involve real-time data processing or urgent decision-making? | Reduced opportunity for human review — elevated risk |
PTA Decision Matrix
| Section A Triggers | Section B Criteria | Section C Factors | Determination |
|---|---|---|---|
| Any trigger met | N/A | N/A | DPIA mandatory |
| No trigger met | 2+ criteria | Any | DPIA strongly recommended |
| No trigger met | 1 criterion | 1+ factors | DPIA recommended |
| No trigger met | 1 criterion | 0 factors | DPO consultation recommended |
| No trigger met | 0 criteria | 1+ factors | Document screening; monitor |
| No trigger met | 0 criteria | 0 factors | DPIA not required |
PTA Documentation Requirements
Regardless of the outcome, every PTA must be documented and retained. The documentation must include:
| Field | Content |
|---|---|
| PTA Reference | Unique identifier (PTA-[ORG]-[YEAR]-[SEQ]) |
| Processing Activity Name | Descriptive name of the processing activity |
| Processing Owner | Name and role of the person responsible |
| Date of PTA | Date the screening was completed |
| Section A Responses | Yes/No for each Art. 35(3) question |
| Section B Responses | Yes/No for each WP248 criterion with justification |
| Section C Responses | Yes/No for each additional factor |
| Determination | DPIA mandatory/recommended/not required |
| DPO Sign-Off | DPO name, date, and whether they agree with the determination |
| Next Steps | Full DPIA initiated / Processing approved without DPIA / Deferred pending further information |
Escalation Process
When a PTA determines that a DPIA is required:
- PTA result is logged in the DPIA register.
- Processing owner is notified that a DPIA must be completed before processing begins.
- DPO assigns a DPIA reference number and initiates the DPIA process.
- Processing must not commence until the DPIA is completed and approved.
When a PTA determines that a DPIA is recommended but not mandatory:
- DPO reviews the PTA and makes a final determination.
- If DPO determines DPIA is needed: escalate as above.
- If DPO determines DPIA is not needed: document the rationale and approve processing.
Common PTA Mistakes
- Skipping the PTA: Proceeding with new processing without any screening — no documentation of the decision not to conduct a DPIA.
- Self-assessment without DPO review: Processing owners conducting PTA without DPO involvement, leading to underestimation of criteria.
- Narrow interpretation of criteria: Interpreting "large scale" or "systematic monitoring" too narrowly to avoid triggering a DPIA.
- No documentation: Conducting a mental screening without documented PTA — cannot demonstrate accountability under Art. 5(2).
- One-time screening: Completing a PTA at project initiation and never revisiting when the processing scope changes.
Sponsors
Details
- Marketplace
- mukul975/Privacy-Data-Protection-Skills
- Source
- plugins/privacy-impact-assessment-skills
- Author
- Mahipal
- Category
- Compliance
- License
- Apache-2.0
- Keywords
- dpia, pia, risk-assessment, gdpr-article-35
More from mukul975/Privacy-Data-Protection-Skills
All 21- privacy-skills-completev1.0.0
- records-of-processing-skillsv1.0.0
- us-state-privacy-skillsv1.0.0
- vendor-privacy-management-skillsv1.0.0
- ai-privacy-governance-skillsv1.0.0
- children-privacy-skillsv1.0.0
- consent-management-skillsv1.0.0
- cookie-consent-skillsv1.0.0
- cross-border-transfers-skillsv1.0.0
- data-breach-response-skillsv1.0.0
- data-classification-skillsv1.0.0
- data-retention-skillsv1.0.0
- data-subject-rights-skillsv1.0.0
- employee-privacy-skillsv1.0.0
- gdpr-compliance-skillsv1.0.0
- global-privacy-regulations-skillsv1.0.0
- healthcare-privacy-skillsv1.0.0
- privacy-audit-skillsv1.0.0
- privacy-by-design-skillsv1.0.0
- privacy-engineering-skillsv1.0.0