The beginning of August sees the coming into application of the EU AI Act (August 2), whilst the EU AI Office also comes into being and has started to hire in earnest.
The EU AI Act will come into practice in a staggered fashion, with prohibition on the unacceptable’ risk category of AI coming into force next February, the general purpose AI model obligations come into effect next August, as does a review of the overall legislation– in addition to the establishment of national AI authorities.
Given this is a new field, we expect the EU to issue delegated acts (acts that help to amend existing legislation) on topics from the technical requirement for general purpose AI to criteria for high-risk AI, in addition to the issue of codes of practice on prohibited AI for instance.
As a side comment, the French police are actively using AI for security reasons at the Olympics and it will be interesting to see if feedback from this otherwise prohibited process feeds back to lawmakers.
At the same time, the EU had been giving some detail on the structure of the new AI Office, which will encompass roles such as a lead scientific officer and an international affairs officer. Other aspects that are of interest are the EU AI Advisory Board, which has had its first meeting and an AI Pact that allows dialogue with corporates.
The Office will be led by Lucilla Sioli, a long-term EU official, and will have five sub-units.
In summary, the EU’s AI infrastructure is taking shape, with hiring across the new AI Office starting to take place. Many of the hires so far are existing officials or from the ranks of MEP research teams, so to an extent there is little industry experience and research/science experts have not yet been brought on board.
