Meta refuses to sign EU’s AI code of practice

Meta refused to sign the European Union Practice code For its AI Act, the general-intention is effective for the AI model suppliers a few weeks before the rules of the block.

“Europe is on the wrong way in AI,” Meta’s chief Global Affairs Officer Joel Kaplan wrote Post At LinkedIn. “We have reviewed the European Commission’s practice code carefully for the general-objective AI (GPAI) models and will not sign it. This code introduces a number of legal uncertainty for model developers, as well as a system that goes much more than the opportunity for AI law.”

EU Practice Code – A voluntary structure published earlier this month – companies help implement processes and systems to comply with block laws for AI control. Among other issues, companies for the code need to supply and update regular documentation about their AI equipment and services; Ban for developers from AI training in pirated materials; And adhere to the owners’ requests for not using their work on their data sets.

Calling for the implementation of the “over-ritch” of the EU law, Kaplan claimed that the law would “developed and deploy border AI models in Europe and European agencies were looking for business at their top.”

Ay Risk -based control for the application of artificial intelligenceThe AI law prohibits some “unacceptable risk” by directing cases such as cognitive behavioral manufacturing or social scoring. The rules also define a set of “high-risky” use, such as biometrics and facial recognition and education and employment in domains. This Act also needs to register the AI systems of the developers and fill the obligations of the risk and standard management.

AI Races like alphabet, meta, Microsoft and Mistral AI who are the world -class technology companies around the world Has been fighting The rules have even called on the European Commission to delay its roll out. However the commission FirmSaying it will not change its timeline.

Friday, published in the EU Guideline AI model suppliers before the rules that will take effect on August 2. These rules will affect the suppliers of “Systemic risky general-object AI models” such as OpenAEA, Ethnic, Google and Meta. Companies that have this national model in the market before August 2 must obey the law by that date.