§ GDPR · AI Act COMPARISON

GDPR and AI Act both apply. Here are the eight specific article-pair collisions Fontvera has extracted from primary text.

No regulation replaces the other. Where they meet — bias data, log retention, automated decisions, technical documentation — the answer is buried in eight cross-references between Regulation (EU) 2016/679 and Regulation (EU) 2024/1689.

Summary

GDPR (Regulation (EU) 2016/679) and the AI Act (Regulation (EU) 2024/1689) both apply, in parallel, to any AI system processing personal data of EU residents. Neither replaces the other. The AI Act explicitly states it is without prejudice to GDPR; GDPR was never amended to defer to the AI Act.

Where they meet, they meet at specific article pairs. Fontvera has mapped eight of these in primary text: four overlaps where the same general control sits in both texts, two hard conflicts where the texts pull in opposite directions, and two structural gaps where neither resolves the boundary. The eight pairs are reproduced below with severity, the entity affected, and the article numbers as extracted from the corpus.

Who this applies to
Any organisation processing personal data of EU residents that also places an AI system on the EU market or uses one — controllers, processors, providers, deployers and joint controllers operating across both regimes simultaneously.
Compliance deadline
GDPR: in force since 25 May 2018. AI Act Article 50 transparency: 2 August 2026. AI Act Annex III high-risk: 2 December 2027 under the Digital Omnibus deal of 7 May 2026, pending Official Journal.
§ Key articles

What the law says

GDPR Article 5
Principles relating to processing — lawfulness, accuracy, storage limitation.
GDPR Article 17
Right to erasure (right to be forgotten).
GDPR Article 22
Right not to be subject to a decision based solely on automated processing.
GDPR Article 24
Responsibility of the controller — technical and organisational measures.
AI Act Article 10
Data and data governance for high-risk AI training, validation and testing sets.
AI Act Article 11
Technical documentation for high-risk AI systems.
AI Act Article 12
Automatic event logging across the lifecycle of a high-risk AI system.
AI Act Article 15
Accuracy, robustness and cybersecurity for high-risk AI systems.
AI Act Article 19
Six-month minimum log retention floor.
AI Act Article 99(2)
Up to €15,000,000 or 3% of worldwide turnover for AI Act provider/deployer breach.
§ Detail

In depth

The eight collisions, mapped

From Fontvera's obligation_crossrefs table, all eight AI Act ↔ GDPR cross-references in one place:

AI ActGDPRTypeSeverityWhat collides
Art 10Art 17ConflictHighAI Act allows processing of special-category data for bias detection where strictly necessary; GDPR right to erasure may require its deletion once retention is no longer necessary. Provider exposure.
Art 19Art 5ConflictMediumAI Act mandates log retention for at least six months; GDPR Article 5(1)(e) storage limitation requires data to be kept no longer than necessary if the logs contain personal data.
Art 12Art 5OverlapLowBoth require logging and record-keeping for accountability — AI Act demands automatic event logging on high-risk systems; GDPR demands records of processing activities under Article 30.
Art 11Art 24OverlapMediumBoth require documentation to demonstrate compliance — AI Act technical documentation under Annex IV; GDPR controller technical and organisational measures.
Art 15Art 24OverlapMediumBoth mandate technical and organisational measures for security and robustness — AI Act focuses on cybersecurity and resilience; GDPR focuses on confidentiality, integrity and availability.
Art 10Art 5OverlapMediumBoth require data quality and accuracy — AI Act mandates bias-free, representative training data; GDPR requires personal data to be accurate and kept up to date.
GapHighNeither regulation clearly defines liability or compliance obligations for deployers of general-purpose AI models when used in high-risk contexts not explicitly covered. Deployer exposure.
GapMediumNo clear guidance on reconciling AI Act transparency in decision-making with GDPR restrictions on disclosing trade secrets or third-party rights. Provider exposure.

The two conflicts you cannot ignore

AI Act Article 10 ↔ GDPR Article 17 — the bias-detection conflict

AI Act Article 10(5) explicitly permits processing of special categories of personal data (race, ethnic origin, political opinion, religious belief, trade union membership, genetic and biometric data, health data, sex life and sexual orientation) for bias detection and correction in high-risk AI systems, where strictly necessary. This is a derogation crafted into the AI Act because conformity assessment under Articles 9–15 cannot be done without it.

GDPR Article 17 grants the data subject the right to erasure once the data is no longer necessary for the original purpose. The two texts collide when the data subject requests erasure of biometric or health data the AI Act provider is using for ongoing bias monitoring. The provider cannot delete the data without breaking AI Act Article 9 risk management; refusing to delete it potentially breaches GDPR Article 17.

The resolution lives in AI Act Article 10(5) read with GDPR Article 9(2)(g) — the "substantial public interest" basis, which the AI Act recognises as a legal basis for the special-category processing. Most providers will need a documented Article 10(5) impact assessment on file to defend retention against an erasure request.

AI Act Article 19 ↔ GDPR Article 5(1)(e) — the log retention conflict

AI Act Article 19 requires providers of high-risk AI systems to keep automatically generated logs for at least six months, and longer where applicable Union or national law mandates. This is a floor.

GDPR Article 5(1)(e) storage limitation requires personal data to be kept "for no longer than is necessary" for the purposes for which it is processed. This is a ceiling.

If the logs contain personal data — which they often will, since they capture inputs, outputs and operator IDs — the operator is squeezed between an AI Act minimum and a GDPR maximum. The reconciliation is documented purpose: the legitimate interest in maintaining AI Act-mandated logs is itself a sufficient purpose under GDPR Article 6, provided the retention period is justified in writing.

Where the two regulations stack at penalty

Both regimes set EU-wide ceilings that apply on top of each other if a single incident breaches both:

An AI system that processes biometric data without a GDPR Article 9 basis and is also classified high-risk under AI Act Annex III could in principle hit both ceilings on the same incident.

Article 22 — the cleanest overlap to resolve

GDPR Article 22 grants data subjects the right not to be subject to a decision based solely on automated processing, including profiling, that produces legal effects or similarly significantly affects them. The AI Act sits on top of this for high-risk Annex III categories — credit scoring (5(b)), insurance pricing (5(c)), employment (4), education (3) — by requiring human oversight under Article 14 and fundamental rights impact assessments under Article 27 for public bodies and Annex III 5(b)/(c) deployers.

The compliance pattern is single-stack: build human oversight into the workflow once, document it under both Article 14 (AI Act) and Article 22(3) (GDPR). The same control clears both audits.

Real numbers Fontvera tracks

Authorities — different bodies, joint enforcement

GDPR is enforced by national Data Protection Authorities (CNIL, BfDI, datatilsynet, IMY, AEPD, Garante, etc.) coordinated by the EDPB. The AI Act is enforced by national market surveillance authorities designated under Article 70, with cross-border coordination by the European AI Office. The two systems already cooperate informally on data-driven AI cases; the AI Office is expected to publish joint guidance with the EDPB on Articles 10, 13, 14 and 22 before Q4 2026.

Where both apply, expect joint investigations. Where the system is high-risk under Annex III and processes personal data, both authorities will arrive at the same audit walkthrough.

What to do before 2 August 2026

  1. Map every AI system that processes personal data against the eight collisions above. The mapping is the deliverable an audit will request.
  2. Document the Article 10(5) basis for special-category bias data in writing, with retention period and deletion criteria. Without it, an Article 17 erasure request defaults to deletion.
  3. Lock log retention in policy. AI Act floor: ≥6 months. GDPR ceiling: documented purpose. Reconcile in a single retention schedule.
  4. Run DPIA and Article 27 FRIA jointly for Annex III 5(b)/(c) deployers — the public-services entities especially. The two documents share most of the analysis.
  5. Treat human oversight as a single control. Build the workflow once, satisfy AI Act Article 14, GDPR Article 22(3), and Article 27 FRIA in one pass.

Run your free AI Act compliance diagnostic

Returns the AI Act classification plus the article list — including which obligations cross-reference back to GDPR.

→ Run the AI Act diagnostic


Cross-regulatory data update

Auto-merged from the Fontvera archetype dataset on 2026-05-12. The sections below are extracted verbatim from the `obligations` and `obligation_crossrefs` tables; the page itself was last reviewed manually before this update.

Summary statistics

Overlaps: 4 · Conflicts: 2 · Gaps: 2

8 article-level crossrefs catalogued between AI Act and GDPR from the Fontvera EU regulatory corpus. Article numbers are verbatim from the underlying obligation_crossrefs table; descriptions are extracted, not paraphrased.

All crossrefs between these regulations

Article (A)Article (B)TypeSeverityDescription
AI Act Art 10GDPR Art 5overlapmedium[entity affected: provider] Both regulations require data quality and accuracy; AI Act mandates bias-free and representative training data, while GDPR requires personal data to be accurate and kept up
AI Act Art 15GDPR Art 24overlapmedium[entity affected: provider] Both regulations mandate the implementation of technical and organizational measures to ensure security and robustness; AI Act focuses on cybersecurity and resilience, whil
AI Act Art 12GDPR Art 5overlaplow[entity affected: provider] Both regulations require logging and record-keeping for accountability; AI Act requires automatic event logging for high-risk systems, while GDPR requires maintaining recor
AI Act Art 11GDPR Art 24overlapmedium[entity affected: provider] Both regulations require documentation to demonstrate compliance; AI Act requires technical documentation for high-risk AI systems, while GDPR requires controllers to imple
AI Act Art 10GDPR Art 17conflicthigh[entity affected: provider] AI Act allows processing special categories of data for bias detection if strictly necessary, which may conflict with GDPR's right to erasure if the data is no longer neces
AI Act Art 19GDPR Art 5conflictmedium[entity affected: provider] AI Act mandates keeping logs for at least six months, which may conflict with GDPR's storage limitation principle requiring data to be kept no longer than necessary if the
AI Act Art ?GDPR Art ?gaphigh[entity affected: deployer] Neither regulation clearly defines the liability or compliance obligations for deployers of general-purpose AI models when used in high-risk contexts not explicitly covered
AI Act Art ?GDPR Art ?gapmedium[entity affected: provider] There is a gap in clear guidance on how to reconcile the AI Act's requirement for transparency in AI decision-making with GDPR's restrictions on disclosing trade secrets or

Conflicts explained

The 2 article-level conflicts between AI Act and GDPR mean a control that satisfies one can pull the wrong way on the other:

Which regulation takes precedence

EU law does not lay down a universal precedence rule between AI Act and GDPR. In practice three resolution approaches apply: lex specialis (the more specific provision wins when both purport to govern the same conduct); regulator guidance (EDPB, EBA, ESMA and the AI Office have all issued joint readings on overlapping articles — check the most recent applicable opinion); and document the choice (when the regulations leave the call to the controller, the audit defence is your written reasoning, not the regulator's silence). Where the corpus surfaces a conflict rather than an overlap, treat that as an escalation path to legal — not a control-design question.

§ Action items

Practical steps

01
Map every AI system processing personal data against the eight AI Act ↔ GDPR collisions in this page. The map is the audit deliverable.
02
Document the Article 10(5) basis for any special-category bias data and align with GDPR Article 9(2)(g) substantial public interest.
03
Lock log retention in policy: ≥6 months under AI Act Article 19, with documented purpose under GDPR Article 5(1)(e).
04
Run DPIA and AI Act Article 27 FRIA jointly for Annex III 5(b) and 5(c) deployers. Share the analysis; do not duplicate it.
05
Treat human oversight as one control that clears AI Act Article 14, GDPR Article 22(3) and the FRIA in a single design.
§ What Fontvera found

Documents in our corpus

imy SE Fetched 2026-06
§ Cross-references

Related Fontvera intelligence

Need a cross-border briefing on this?
Search Fontvera ↵ Run the AI Act diagnostic
AI Act Article 50 transparency
50 days
until 2026-08-02, when Article 50 transparency obligations apply (unchanged). Annex III high-risk obligations move provisionally to 2 December 2027 under the Digital Omnibus agreement of 7 May 2026, pending formal adoption.
Preparing for 2 August 2026? Read the EU AI Act August 2026 deadline requirements checklist.