Meta has announced that it will not sign a new voluntary European Union Code of Practice aimed at regulating general-purpose artificial intelligence (GPAI) models, Engadget reports. The company cited the document's legal ambiguity and criticism of excessive regulatory intervention as reasons for the decision.
Meta's chief international affairs officer, Joel Kaplan, said the code "goes beyond" what is required by the AI Act, adding that its passage creates additional risks for developers and could stifle innovation in the industry.
The code, published on July 10, contains requirements for companies, including a ban on using pirated materials to train AI models, an obligation to consider authors' requests to exclude their works from training data, and regular updates of technical documentation on AI functions.
Despite the voluntary nature of the Code, companies that refuse to sign it could face increased scrutiny from regulators. European Commission Digital Commissioner Thomas Rainier stressed that such developers would have to prove their compliance with the standards in other ways, which could complicate their legal situation.
Violations of the AI Act are subject to fines of up to 7% of a company's annual turnover. For developers of advanced models, this threshold is 3%.