The GDPR and AI
In effect since 2018, the GDPR is the global baseline for data protection, deliberately technology-agnostic so it can evolve alongside AI. Three provisions intersect most with AI: Article 22 (automated decisions), Article 35 (DPIAs) and Recital 26 (anonymisation).
In effect since 2018, the GDPR is the global baseline for data protection, deliberately technology-agnostic so it can evolve alongside AI.
AI systems consume extensive data, so when that includes personal data they fall within the GDPR's scope → collection, use, protection and control duties, including supporting the right to delete. The underpinning principles → lawfulness, fairness and transparency · purpose limitation · data minimisation · accuracy · storage limitation · integrity and confidentiality · accountability.
The three provisions that intersect with AI:
- Article 22 (automated decision-making) → requires a review process for decisions with adverse or material impacts on individuals.
- Article 35 (DPIAs) → mandates data protection impact assessments for high-risk or significant processing activities.
- Recital 26 (anonymisation) → emphasises pseudonymisation and anonymisation techniques to safeguard personal data in AI development.