Meta has publicly refused to sign the European Union’s voluntary AI Code of Practice, calling the framework an overreach and warning it could stifle innovation.

The voluntary code is designed to guide companies in complying with the EU’s sweeping AI Act, which came into effect in June 2024.

It includes transparency and copyright requirements, urging developers to disclose information about training data and model capabilities.

Meta’s Global Affairs Chief Joel Kaplan stated on LinkedIn that “Europe is heading down the wrong path on AI,” and argued the code “introduces legal uncertainties” and goes “far beyond the scope” of the AI Act.

His comments align Meta with a group of 45 European firms, including Airbus, ASML and Mistral AI, who recently urged the European Commission to pause the AI Act’s rollout for two years to avoid hampering innovation.

Meanwhile, Microsoft has signaled support for the framework.

Microsoft President Brad Smith told Reuters the company is “likely” to sign the code, pending review, adding they welcome direct engagement with EU regulators.

EU digital spokesperson Thomas Regnier warned that companies opting out of the code will face heightened regulatory scrutiny and must prove compliance through alternative measures.

With enforcement on general-purpose AI models set to begin next month, the standoff underscores growing friction between US tech giants and EU lawmakers.

Breaches of the AI Act could result in fines of up to 7% of global revenue.