EDPB and EDPS adopt Joint Opinion on the Digital Omnibus on AI
The European Data Protection Board (EDPB) and the European Data Protection Supervisor (EDPS) have issued a Joint Opinion on the European Commission’s Proposal for a “Digital Omnibus on AI,” which aims to simplify implementation of certain harmonised rules under the AI Act. While welcoming efforts to reduce administrative burdens and promote innovation, the supervisory bodies stress that simplification must not weaken protections for fundamental rights or diminish accountability for AI providers and deployers.
Key concerns focus on proposed extensions and deletions that could undermine data protection safeguards. The Opinion warns against broadly permitting processing of special categories of personal data (for example ethnicity or health) for bias detection and correction, recommending instead narrow, well-justified use only where the risk of serious adverse effects from bias is demonstrable. The EDPB and EDPS also oppose removing the obligation to register AI systems currently listed as high-risk, arguing that such a deletion would erode public scrutiny and create incentives for providers to misclassify systems to avoid oversight.
The supervisory bodies support EU-level AI regulatory sandboxes as a tool for innovation and SME support but insist that Data Protection Authorities (DPAs) must be directly involved in supervision of data processing within sandboxes. They recommend that the EDPB have an advisory role and observer status at the European Artificial Intelligence Board to secure consistency, and that the AI Office’s supervisory remit—particularly over systems based on general-purpose models—be clearly delimited so it does not overlap with the EDPS’s independent supervision of Union institutions’ AI use.
Finally, the Joint Opinion urges that cooperation with Market Surveillance Authorities be clarified to preserve DPA independence and powers, and recommends keeping mandatory AI literacy duties for providers and deployers rather than shifting core responsibilities to the Commission or Member States. The EDPB and EDPS also caution against postponing core obligations for high-risk AI systems, urging co-legislators to minimise delays for transparency and other essential requirements given the fast-evolving AI landscape.
