Meta refuses to sign up to EU Artificial Intelligence Bill just weeks before EU general AI model provider rules come into effectCode of Conduct. Joel Kaplan, chief global affairs officer at Meta, leads theBritain wrote: "Europe is on the wrong track in the field of artificial intelligence. We have carefully reviewed the European Commission's code of conduct for general artificial intelligence (GPAI) models, and Meta will not sign it. The code creates a lot of legal uncertainty for model developers and its measures go well beyond the scope of the Artificial Intelligence Act."
The EU’s Code of Conduct – a voluntary framework published earlier this month to help businesses implement processes and systems to comply with EU legislation on the regulation of artificial intelligence. Among other things, the guidelines require companies to provide and regularly update documentation for their AI tools and services; prohibit developers from using pirated content to train AI; and comply with requests from content owners not to use their work in their data sets.
Kaplan called the EU's implementation of the legislation "overreach" and claimed that the law would "restrict the development and deployment of cutting-edge AI models in Europe and prevent European companies from operating on the basis of these models."
The Artificial Intelligence Act is a risk-based regulation of artificial intelligence applications that outright bans certain “unacceptably risky” use cases, such as cognitive behavioral manipulation or social scoring. The rules also define a range of "high-risk" uses, such as biometrics and facial recognition, as well as applications in areas such as education and employment. The bill also requires developers to register artificial intelligence systems and meet risk and quality management obligations.
Tech companies around the world, including those at the forefront of the race for artificial intelligence such as Alphabet, Meta, Microsoft and Mistral AI, have been fighting the rules and even urged the European Commission to delay their rollout. But the European Commission stood firm and said it would not change the timetable.
Also on Friday, the European Union published guidance for providers of artificial intelligence models, which will come into effect on August 2. The rules will affect providers of "general artificial intelligence models with systemic risks" such as OpenAI, Anthropic, Google and Meta. Companies that put such models on the market before August 2 have until August 2, 2027 to comply with the regulation.