Council of the European Union Publishes Key Priorities for the Second EU AI Act Trilogue

The Council of the European Union has published a summary outlining the provisional agreements, pending items, and future priorities discussed during the second trilogue on the sweeping EU AI Act.

The past few months have seen significant progress made with the EU AI Act, which seeks to set the global gold standard for AI regulation with its risk-based approach.

Following the adoption of the then draft General Approach by the Council of the European Union on 6 December 2022, leading Parliamentary Committees accepted the Adopted Text on 11 May 2023 ahead of the European Parliament vote on 14 June 2023, where the text was passed by majority vote, paving the way for trilogues to commence.

The first trilogue took place on 14 June after the vote, with the second taking place on 18 July following technical consultations and an all-day meeting of the Working Party on Telecommunications and Information Society on 5 July 2023, where delegations outlined their priorities.

So far, discussions have revolved around:

• Obligations of providers and users/deployers of high-risk AI systems and other
  parties (Articles 16-27).
• Notifying authorities and notified bodies (Articles 30-39).
• Standards, conformity assessment, certificates, registration (Articles 40-50).

Seeking to make the legislative process as transparent as possible, the Council of the European Union has made available a (partially redacted) summary of the key changes made to these articles during the technical meetings. The document indicates the parts of the text with provisional agreement at the technical level, those pending provisional agreement, and parts of the text to be reviewed at a later stage, setting out the priorities for the second trilogue.

Trilogues are expected to wrap up by the end of 2023, with the final text likely to be published shortly after. The EU AI Act is still very much evolving, but each publication is one step closer to AI used in the EU being made safer.

The sweeping EU AI Act is set to impose a raft of obligations on organisations within the EU as well as non-EU companies who operate AI systems within its borders. To find out how Holistic AI can help you manage the risks of your AI systems, schedule a call with our expert team.