The AI Act - what do we know so far?

After a 36-hour negotiating marathon, a provisional agreement on the EU AI Act was struck between EU policymakers on Friday (8th). The provisional agreement outlines that the AI Act should apply two years after its entry into force, with some exceptions for specific provisions.

Compared to the initial Commission proposal, the main new elements of the provisional agreement are as follows:

• New rules on general purpose AI systems and foundation models

New provisions have been added to take into account general purpose AI, and its potential integration into high-risk systems.

The provisional agreement provides that foundation models must comply with specific transparency obligations before they are placed in the market. A stricter regime was introduced for ‘high impact' foundation models.

• A revised system of governance with some enforcement powers at EU level

An AI Office within the Commission will be tasked to oversee these most advanced AI models, contribute to fostering standards and testing practices, and enforce the common rules in all member states.

The AI Board, which would comprise member states' representatives, will remain as a coordination platform and an advisory body to the Commission and will give an important role to Member States on the implementation of the regulation, including the design of codes of practice for foundation models. An advisory forum for stakeholders, such as industry representatives, will be set up to provide technical expertise to the AI Board.

• Extension of the list of prohibitions but with the possibility to use remote biometric identification by law enforcement authorities in public spaces, subject to safeguards

Some narrow law enforcement exceptions, namely to prevent terrorist attacks or locate the victims or suspects of a pre-defined list of serious crimes, are expected instead of a general ban on using real-time remote biometric identification.

• An obligation for deployers of high-risk AI systems to conduct a fundamental rights impact assessment prior to putting an AI system into use.

Next steps

Work will continue at a technical level over the coming weeks to finalise the details of the new regulation. The presidency will submit the compromise text to the member states' representatives for endorsement once this work has been concluded. The entire text will need to be confirmed by both institutions and undergo legal-linguistic revision before formal adoption by the co-legislators. This translation exercise is likely to take some time.

techUK await the final text before determining the implications of the Act and what it means for our members. Although it's fair to say that the hard work has only just begun, and details as to how the Act will be effectively delegated and implemented by member states is still to be determined. Questions also still remain as to how this approach will be interoperable with the UK's planned approach. We look forward to further detail on this imminently when government outlines it's response to the UK AI Whitepaper.