Overview

• Meta became the first major AI provider to reject the EU’s voluntary Code of Practice, announcing its decision on July 18 less than a month after the code’s July 10 release.
• The company’s chief global affairs officer, Joel Kaplan, criticized the framework for exceeding the AI Act’s scope and introducing legal uncertainties that could hinder frontier AI deployment.
• The European Commission says firms signing the code will benefit from clearer legal guidance and lighter enforcement under the AI Act when its general-purpose AI rules begin on August 2.
• Mistral and OpenAI have joined as early signatories, while more than 40 European firms in July called for a delay to new AI Act obligations.
• The code lays out requirements for continuous documentation, bans on using pirated content and protocols to honor data rights, aiming to operationalize the AI Act’s risk-based compliance regime.