Meta refuses to sign EU AI agreement, citing overreach and legal uncertainty
&imwidth=600&imheight=450&format=webp&quality=medium "Meta refuses to sign EU AI agreement, citing overreach and legal uncertainty")
People walk behind a logo of Meta Platforms company, during a conference in Mumbai, India, September 20, 2023. Photograph: (Reuters)
Several companies, including OpenAI and France-based Mistral, have already committed to the voluntary framework, which requires developers to enhance transparency, adhere to EU copyright law, and disclose the data used to train their models.
Meta Platforms has announced it will not sign the European Commission’s new Code of Practice for general-purpose artificial intelligence (AI), criticising it as an overreach that threatens innovation and introduces legal uncertainty. The voluntary code, released earlier this month, is designed to help companies comply with the EU’s sweeping AI Act, which came into effect in June 2024 and will begin regulating general-purpose AI models starting August 2. Meta’s Chief Global Affairs Officer Joel Kaplan voiced the company’s opposition in a post on LinkedIn on July 18, stating, “Europe is heading down the wrong path on AI. This code introduces a number of legal uncertainties for model developers, as well as measures that go far beyond the scope of the AI Act.”
Kaplan argued that the framework could “stunt” the development of frontier AI models and hinder the growth of European businesses looking to build on such technologies. He added that Meta’s stance aligns with a growing number of European firms, including Airbus, ASML, and Mistral AI, that recently signed an open letter urging the EU to delay implementation of the AI Act by two years to avoid stifling innovation.
Microsoft likely to sign; others already onboard
In contrast, Microsoft President Brad Smith signalled that the company is inclined to sign the code. “I think it's likely we will sign,” Smith told Reuters, adding that Microsoft welcomes the direct engagement between the EU’s AI Office and industry stakeholders.
Several companies, including OpenAI and France-based Mistral, have already committed to the voluntary framework, which requires developers to enhance transparency, adhere to EU copyright law, and disclose the data used to train their models. The code, developed with input from over 1,000 stakeholders, is intended to offer legal clarity and reduce compliance burdens once it is formally endorsed by EU member states.
Tensions reflect transatlantic divide
Trending Stories
&im=FitAndFill=(700,400))
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
)
)
)
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
)
)
&im=FitAndFill=(700,400))
)
)
)
Meta’s rejection underscores growing tension between US tech giants and European regulators over digital governance. The EU argues its regulatory approach ensures transparency and public safety, while companies like Meta fear the rules may inhibit competitiveness and technological leadership in AI. Violations of the AI Act can lead to fines of up to 7 per cent of a company’s annual global revenue, adding significant weight to compliance decisions. Despite the voluntary nature of the code, it is widely seen as a blueprint for future regulatory enforcement across the bloc.
(With inputs from agencies)