Google signs the EU code of conduct on artificial intelligence, unlike Meta

These European recommendations take place as Grok, Elon Musk’s AI recently hit the headlines by relaying extremist and abusive remarks.

Google will sign the European Union code of conduct (EU), baptized “Ia act”On the supervision of artificial intelligence models (IA), Kent Walker, Group World Affairs President, announced on Wednesday July 30 on Wednesday July 30. OPENAI, creator of Chatgpt, and the French start-up Mistral had already announced the signing of this code of conduct, while Meta, parent company of Facebook and Instagram, said that it would not.

Published on July 10 and implemented from August 2, these European recommendations, thoughts for AI models for general use such as Chatgpt, Grok of the X or Gemini platform of Google, focus on copyright issues. The EU also calls to exclude from AI regularly hacked sites and abusive or violent remarks so that this information is not taken up by the algorithms of the signatories.

In addition, there are reinforced requirements to which high -risk applications will be submitted before any marketing authorization in Europe, by 2026. The latter will concern in particular the platforms dedicated to education, human resources or even the maintenance of order.

Future recommendations that annoy digital giants, some of which call to postpone the law. Wednesday, Google thus mentioned European rules which “risk slowing the development of AI in Europe”.

And what “Code of good practices” is not binding, signatory companies will however benefit from a “Reduced administrative load” When it is a question of proving that they comply well with European legislation on AI, promises the European Commission.