Originally scheduled for Friday 2 May, the final version of the ‘General-Purpose AI Code of Practice’ is not expected to see the light of day until early summer.
Plagued by numerous criticisms (see EUROPE 13610/3) and an intense lobbying campaign by the digital giants (see EUROPE 13613/25) which has been denounced by several European associations and NGOs (see EUROPE 13631/14), the drafting of the Code remains controversial and far from unanimous.
The text, which is intended to guide artificial intelligence providers in their compliance with the relevant regulations (AI Act), was recently condemned by the US government’s mission to the EU, which went so far as to call for its withdrawal (see EUROPE 13629/7).
The Code is due to be published “before August”, according to several sources. The exact date – which has not been communicated by the European Commission to stakeholders – should however coincide with the entry into force of certain provisions of the AI Act regarding sanctions.
The new voluntary nature of certain measures that are outlined in the Code, such as those aimed at preventing human rights violations, privacy violations, has been widely denounced by many players in the sector, as well as by the MEPs involved in drafting it (see EUROPE 13608/6).
On the other hand, several associations from the European digital sector have denounced the “additional administrative burden” that the Code might represent.
Despite these major differences of opinion, several experts involved in the drafting process believe that nevertheless “some kind of common ground is in sight”. (Original version in French by Isalia Stieffatre)