- Adaptation of DPA and T&Cs - Refer to privacy statement as DPA, not a privacy statement - Startup of enforcing signed DPA and T&Cs - Adaptation of eveai_chat_client to ensure we retrieve correct DPA & T&Cs
20 KiB
Data Protection Impact Assessment (DPIA) Template
Ask Eve AI
Date of Assessment: [Date]
Assessed By: [Name, Role]
Review Date: [Date - recommend annual review]
1. Executive Summary
| Field | Details |
|---|---|
| Processing Activity Name | [e.g., "Job Candidate Assessment Specialist"] |
| Brief Description | [1-2 sentence summary] |
| Risk Level | ☐ Low ☐ Medium ☐ High |
| DPIA Required? | ☐ Yes ☐ No |
| Status | ☐ Draft ☐ Under Review ☐ Approved ☐ Requires Revision |
2. Description of the Processing
2.1 Nature of the Processing
What Personal Data will be processed?
- Contact information (name, email, phone)
- Identification data (ID numbers, passport)
- Professional data (CV, work history, qualifications)
- Assessment results or scores
- Communication records
- Behavioral data (how users interact with the system)
- Technical data (IP addresses, device information)
- Other: _______________
Categories of Data Subjects:
- Job applicants/candidates
- Employees
- Customers
- End users/consumers
- Other: _______________
Volume of Data Subjects:
- < 100
- 100-1,000
- 1,000-10,000
- > 10,000
2.2 Scope of the Processing
What is the purpose of the processing?
[Describe the specific business purpose, e.g., "To assess job candidates' suitability for specific roles by analyzing their responses to standardized questions"]
How will the data be collected?
- Directly from data subjects (forms, interviews)
- From third parties (recruiters, references)
- Automated collection (web forms, chatbots)
- Other: _______________
Where will data be stored?
- EU (specify: France - Scaleway)
- Non-EU (specify and justify): _______________
2.3 Context of the Processing
Is this processing new or existing?
- New processing activity
- Modification of existing processing
- Existing processing (periodic review)
Who has access to the Personal Data?
- Ask Eve AI employees (specify roles): _______________
- Customer/Tenant employees
- Partners (specify): _______________
- Sub-Processors (list): _______________
- Other: _______________
How long will data be retained?
[Specify retention period and justification, e.g., "Candidate data retained for 12 months to comply with recruitment record-keeping requirements"]
3. Necessity and Proportionality Assessment
3.1 Lawful Basis
What is the lawful basis for processing? (Article 6 GDPR)
- Consent - Data subject has given explicit consent
- Contract - Processing necessary for contract performance
- Legal obligation - Required by law
- Vital interests - Necessary to protect someone's life
- Public task - Performing a public interest task
- Legitimate interests - Necessary for legitimate interests (requires balancing test)
Justification:
[Explain why this lawful basis applies]
3.2 Special Categories of Data (if applicable)
Does the processing involve special categories of data? (Article 9 GDPR)
- No
- Yes - racial or ethnic origin
- Yes - political opinions
- Yes - religious or philosophical beliefs
- Yes - trade union membership
- Yes - genetic data
- Yes - biometric data for identification
- Yes - health data
- Yes - sex life or sexual orientation data
If yes, what is the additional lawful basis?
[Article 9(2) provides specific conditions - specify which applies]
3.3 Automated Decision-Making
Does the processing involve automated decision-making or profiling?
- No
- Yes - automated decision-making WITH human oversight
- Yes - fully automated decision-making (no human intervention)
If yes:
Does it produce legal effects or similarly significant effects?
- No
- Yes (explain): _______________
What safeguards are in place?
- Right to obtain human intervention
- Right to express point of view
- Right to contest the decision
- Regular accuracy reviews
- Transparency about logic involved
- Other: _______________
3.4 Necessity Test
Is the processing necessary to achieve the stated purpose?
☐ Yes ☐ No
Justification:
[Explain why this specific processing is necessary and whether less intrusive alternatives were considered]
Could the purpose be achieved with less data or through other means?
☐ Yes (explain why not pursued): _______________
☐ No
3.5 Proportionality Test
Is the processing proportionate to the purpose?
☐ Yes ☐ No
Data Minimization:
- Are you collecting only the minimum data necessary? ☐ Yes ☐ No
- Have you considered pseudonymization or anonymization? ☐ Yes ☐ No ☐ N/A
- Can data be aggregated instead of individual records? ☐ Yes ☐ No ☐ N/A
Storage Limitation:
- Is the retention period justified and documented? ☐ Yes ☐ No
- Is there an automated deletion process? ☐ Yes ☐ No ☐ Planned
4. Stakeholder Consultation
4.1 Data Subject Consultation
Have data subjects been consulted about this processing?
☐ Yes ☐ No ☐ Not required
If yes, how were they consulted?
[Describe consultation method: surveys, focus groups, user research, etc.]
Key concerns raised by data subjects:
[List any concerns and how they were addressed]
4.2 DPO or Security Contact Consultation
Has the DPO or security contact been consulted?
☐ Yes ☐ No ☐ N/A (no formal DPO)
Comments from DPO/Security Contact:
[Record any recommendations or concerns]
5. Risk Assessment
5.1 Risk Identification
For each risk, assess:
- Likelihood: Negligible / Low / Medium / High
- Severity: Negligible / Low / Medium / High
- Overall Risk: Low / Medium / High / Very High
Risk 1: Unauthorized Access or Data Breach
Description: Personal data could be accessed by unauthorized parties due to security vulnerabilities.
| Assessment | Rating |
|---|---|
| Likelihood | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Severity (if occurs) | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Overall Risk | ☐ Low ☐ Medium ☐ High ☐ Very High |
Risk 2: Discrimination or Bias in Automated Decisions
Description: Automated processing could result in discriminatory outcomes or unfair treatment.
| Assessment | Rating |
|---|---|
| Likelihood | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Severity (if occurs) | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Overall Risk | ☐ Low ☐ Medium ☐ High ☐ Very High |
Risk 3: Lack of Transparency
Description: Data subjects may not understand how their data is processed or decisions are made.
| Assessment | Rating |
|---|---|
| Likelihood | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Severity (if occurs) | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Overall Risk | ☐ Low ☐ Medium ☐ High ☐ Very High |
Risk 4: Inability to Exercise Data Subject Rights
Description: Data subjects may have difficulty exercising their rights (access, erasure, portability, etc.).
| Assessment | Rating |
|---|---|
| Likelihood | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Severity (if occurs) | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Overall Risk | ☐ Low ☐ Medium ☐ High ☐ Very High |
Risk 5: Data Quality Issues
Description: Inaccurate or outdated data could lead to incorrect decisions or outcomes.
| Assessment | Rating |
|---|---|
| Likelihood | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Severity (if occurs) | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Overall Risk | ☐ Low ☐ Medium ☐ High ☐ Very High |
Risk 6: Function Creep / Scope Expansion
Description: Data collected for one purpose could be used for other purposes without consent.
| Assessment | Rating |
|---|---|
| Likelihood | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Severity (if occurs) | ☐ Negligible ☐ Low ☐ Medium ☐ High |
| Overall Risk | ☐ Low ☐ Medium ☐ High ☐ Very High |
Additional Risks:
[Add any processing-specific risks]
6. Mitigation Measures
For each identified risk, document mitigation measures:
Risk 1: Unauthorized Access or Data Breach
Mitigation Measures:
- Encryption in transit (TLS 1.2+)
- Encryption at rest
- Multi-factor authentication
- Access controls (RBAC)
- Regular security audits
- WAF and DDoS protection (Bunny.net Shield)
- Multi-tenant data isolation
- Regular security training
- Incident response plan
- Other: _______________
Residual Risk After Mitigation: ☐ Low ☐ Medium ☐ High ☐ Very High
Risk 2: Discrimination or Bias in Automated Decisions
Mitigation Measures:
- Regular bias testing of AI models
- Diverse training data sets
- Human review of automated decisions
- Clear criteria for decision-making
- Right to contest decisions
- Transparency about decision logic
- Regular fairness audits
- Monitoring of outcomes by demographic groups
- Ability to request explanation
- Other: _______________
Residual Risk After Mitigation: ☐ Low ☐ Medium ☐ High ☐ Very High
Risk 3: Lack of Transparency
Mitigation Measures:
- Clear Privacy Policy explaining processing
- Explicit consent mechanisms
- Plain language explanations
- Information provided before data collection
- Explanation of automated decision logic
- Contact information for questions
- Regular communication with data subjects
- Privacy-by-design approach (anonymous until consent)
- Other: _______________
Residual Risk After Mitigation: ☐ Low ☐ Medium ☐ High ☐ Very High
Risk 4: Inability to Exercise Data Subject Rights
Mitigation Measures:
- Clear procedures for rights requests
- Multiple request channels (email, helpdesk)
- 30-day response timeframe
- Technical capability to extract data
- Data portability in standard formats
- Secure deletion processes
- Account disabling/restriction capability
- Identity verification procedures
- Other: _______________
Residual Risk After Mitigation: ☐ Low ☐ Medium ☐ High ☐ Very High
Risk 5: Data Quality Issues
Mitigation Measures:
- Data validation on input
- Regular data accuracy reviews
- Ability for data subjects to correct errors
- Clear data update procedures
- Data quality monitoring
- Source verification for third-party data
- Archiving of outdated data
- Other: _______________
Residual Risk After Mitigation: ☐ Low ☐ Medium ☐ High ☐ Very High
Risk 6: Function Creep / Scope Expansion
Mitigation Measures:
- Documented purpose limitation
- Access controls preventing unauthorized use
- Regular compliance audits
- Privacy Policy clearly states purposes
- Consent required for new purposes
- Technical controls preventing misuse
- Staff training on data protection
- Other: _______________
Residual Risk After Mitigation: ☐ Low ☐ Medium ☐ High ☐ Very High
Additional Mitigation Measures
[Document any additional mitigation measures not covered above]
7. Data Subject Rights Implementation
How will you ensure data subjects can exercise their rights?
Right of Access (Article 15)
- Procedure documented
- Technical capability implemented
- Response within 30 days
- Method: _______________
Right to Rectification (Article 16)
- Procedure documented
- Technical capability implemented
- Response within 30 days
- Method: _______________
Right to Erasure (Article 17)
- Procedure documented
- Technical capability implemented
- Response within 30 days
- Method: _______________
- Limitations: _______________
Right to Restriction (Article 18)
- Procedure documented
- Technical capability implemented (account disabling)
- Response within 30 days
Right to Data Portability (Article 20)
- Procedure documented
- Technical capability implemented
- Export format: JSON / CSV / XML / Other: _______________
Right to Object (Article 21)
- Procedure documented
- Opt-out mechanisms implemented
- Clear in Privacy Policy
Rights Related to Automated Decision-Making (Article 22)
- Human intervention available
- Explanation of logic provided
- Right to contest implemented
- Documented in Privacy Policy
8. Privacy by Design and Default
Privacy Enhancing Technologies Implemented:
- Data minimization (collect only necessary data)
- Pseudonymization (where applicable)
- Anonymization (where applicable)
- Anonymous interaction until consent (privacy-by-design)
- Encryption (in transit and at rest)
- Access controls and authentication
- Audit logging
- Secure deletion
- Data isolation (multi-tenant architecture)
- Other: _______________
Default Settings:
- Most privacy-protective settings by default
- Opt-in (not opt-out) for non-essential processing
- Clear consent mechanisms before data collection
- Limited data sharing by default
9. Compliance with Principles
For each GDPR principle, confirm compliance:
Lawfulness, Fairness, Transparency (Article 5(1)(a))
- Lawful basis identified and documented
- Processing is fair and transparent
- Privacy Policy clearly explains processing
- Evidence: _______________
Purpose Limitation (Article 5(1)(b))
- Specific purposes documented
- Data not used for incompatible purposes
- New purposes require new consent/legal basis
- Evidence: _______________
Data Minimization (Article 5(1)(c))
- Only necessary data collected
- Regular review of data collected
- Excess data not retained
- Evidence: _______________
Accuracy (Article 5(1)(d))
- Mechanisms to ensure data accuracy
- Ability to correct inaccurate data
- Regular data quality reviews
- Evidence: _______________
Storage Limitation (Article 5(1)(e))
- Retention periods defined and documented
- Automated deletion where appropriate
- Justification for retention documented
- Evidence: _______________
Integrity and Confidentiality (Article 5(1)(f))
- Appropriate security measures implemented
- Protection against unauthorized access
- Encryption and access controls in place
- Evidence: See Annex 2 of DPA
Accountability (Article 5(2))
- Documentation of compliance measures
- Records of processing activities maintained
- DPIA conducted and documented
- DPA in place with processors
- Evidence: This DPIA, DPA with customers
10. International Transfers
Does this processing involve transfer to third countries?
☐ No - all processing within EU
☐ Yes (complete below)
If yes:
Country/Region: _______________
Transfer Mechanism:
- Adequacy decision (Article 45)
- Standard Contractual Clauses (Article 46)
- Binding Corporate Rules (Article 47)
- Other: _______________
Transfer Impact Assessment Completed? ☐ Yes ☐ No
Additional Safeguards:
[Document supplementary measures to ensure adequate protection]
11. Documentation and Records
Documentation Maintained:
- This DPIA
- Privacy Policy
- Data Processing Agreement
- Consent records (if applicable)
- Records of processing activities (Article 30)
- Data breach register
- Data Subject rights request log
- Staff training records
- Sub-processor agreements
Record of Processing Activities (Article 30) Completed?
☐ Yes ☐ No ☐ In Progress
12. Outcomes and Recommendations
12.1 Overall Risk Assessment
After implementing mitigation measures, what is the residual risk level?
☐ Low - processing can proceed
☐ Medium - additional measures recommended
☐ High - significant concerns, consult DPO/legal counsel
☐ Very High - processing should not proceed without major changes
12.2 Recommendations
Recommended Actions Before Processing Begins:
- [Action item 1]
- [Action item 2]
- [Action item 3]
Recommended Monitoring/Review Activities:
- [Monitoring item 1]
- [Monitoring item 2]
- [Monitoring item 3]
12.3 Consultation with Supervisory Authority
Is consultation with supervisory authority required?
☐ No - residual risk is acceptable
☐ Yes - high residual risk remains despite mitigation (Article 36)
If yes, when will consultation occur? _______________
12.4 Sign-Off
DPIA Completed By:
Name: _______________
Role: _______________
Date: _______________
Signature: _______________
Reviewed and Approved By:
Name: _______________
Role: _______________
Date: _______________
Signature: _______________
Next Review Date: _______________
(Recommend annual review or when significant changes occur)
Appendix A: Completed Example - Job Candidate Assessment
This appendix provides a completed example for reference.
Example: Job Candidate Assessment Specialist
Processing Activity: AI-powered job candidate assessment tool
Personal Data Processed:
- Assessment responses (text)
- Communication records (chatbot interactions)
- Contact information (name, email) - collected AFTER assessment with consent
- Assessment scores/results
Purpose: To assess candidates' suitability for job roles based on their responses to standardized questions
Lawful Basis:
- Consent (candidates explicitly consent before providing contact information)
- Contract (processing necessary to take steps at request of data subject prior to entering into contract)
Automated Decision-Making: Yes, with human oversight. Candidates are assessed by AI, but:
- Contact information only collected AFTER positive assessment
- Human recruiter makes final hiring decisions
- Candidates can restart assessment at any time
- Candidates informed about AI assessment before beginning
Key Risks Identified:
- Bias/discrimination in assessment algorithms - MEDIUM risk
- Lack of transparency about assessment criteria - MEDIUM risk
- Data breach exposing candidate information - LOW risk (after mitigation)
Key Mitigation Measures:
- Anonymous assessment until consent obtained
- Clear explanation of assessment process
- Right to contest results
- Human review of all final decisions
- Regular bias testing of algorithms
- Strong technical security measures (encryption, access controls)
- 12-month retention period with secure deletion
Residual Risk: LOW - processing can proceed
Special Considerations:
- Candidates must be informed about automated decision-making
- Privacy Policy must explain assessment logic
- Contact information collected only after explicit consent
- Right to human intervention clearly communicated
Appendix B: Resources and References
GDPR Articles Referenced:
- Article 5: Principles relating to processing
- Article 6: Lawfulness of processing
- Article 9: Special categories of data
- Article 13-14: Information to be provided
- Article 15-22: Data subject rights
- Article 22: Automated decision-making
- Article 28: Processor obligations
- Article 30: Records of processing activities
- Article 33-34: Data breach notification
- Article 35: Data Protection Impact Assessment
- Article 36: Prior consultation with supervisory authority
- Article 45-46: International transfers
Additional Guidance:
- WP29 Guidelines on DPIAs (WP 248)
- WP29 Guidelines on Automated Decision-Making (WP 251)
- ICO DPIA Guidance
- EDPB Guidelines on processing personal data for scientific research
- Belgian DPA Guidance (https://www.gegevensbeschermingsautoriteit.be)
Internal Documents:
- Ask Eve AI Data Protection Agreement
- Ask Eve AI Privacy Policy
- Technical and Organizational Measures (DPA Annex 2)
End of DPIA Template