Through the voice of its head of international affairs, Meta (the parent company of Facebook and Instagram) declared its opposition to the signing of the future Artificial Intelligence (AI) Code of Practice, at Meta’s EU Innovation Day 2025 in Brussels on Tuesday 4 February.
The AI Code of Practice, the final version of which is intended to guide the development and deployment of future AI models in the EU to ensure their reliability, is “unworkable and infeasible”, said Joel Kaplan, as reported by Bloomberg media.
The Code adds “unnecessary hurdles for Western open-source AI models”, he claimed.
He is not alone in criticising the drafting of the Code. Eight lobbies from the digital sector sent a statement to the European Commission at the beginning of January, in which they argued that the Code of Practice would add “additional requirements of a political or legislative nature not provided for in the scope of the AI Regulation” (see EUROPE 13560/5).
The head of Meta also spoke about the roll-out of Community Notes, the new collaborative feature that is set to replace the work of fact-checkers (see EUROPE 13553/5).
It is due to be launched in the United States in the “next few months”, then gradually from 2026 for the rest of the world.
The tool, copied from X’s ‘Notes’, is clearly struggling to convince the European institutions of its ability to meet requirements in terms of content moderation and the fight against online disinformation (see EUROPE 13564/4). (Original version in French by Isalia Stieffatre)