EU artificial intelligence regulatory framework now divided: Microsoft Corporation (MSFT.US) plans to join the code of conduct, but Meta (META.US) explicitly refuses.

The President of the European Union revealed on Friday that the EU may adopt its Artificial Intelligence Code of Conduct to assist companies in complying with the AI Act, while Meta has explicitly refused to participate in this voluntary framework. The code of conduct, drafted by 13 independent experts, aims to provide legal certainty for signatories, requiring participating companies to disclose summaries of training data for general AI models and develop policies in compliance with EU copyright law. As a supplementary measure to the AI Act effective in June 2024, this regulatory framework will apply to thousands of companies including Alphabet Inc. Class C, Meta, OpenAI, Anthropic, Mistral, and others. Brad Smith, President of Microsoft Corporation, stated in an interview: "We are likely to sign the code of conduct, but we need to carefully review the contents of the document. We welcome the mechanism for direct communication between the EU AI office and companies." Joel Kaplan, Chief Global Affairs Officer at Meta, reiterated his opposition to the code on LinkedIn, stating: "The code contains multiple legal uncertainties, and its measures go beyond the authorization scope of the AI Act." He emphasized that Meta, along with 45 European companies, shares common concerns and believes that excessive regulatory expansion will hinder the development and application of cutting-edge AI models in Europe, thus not conducive to local businesses innovating based on this technology. Currently, OpenAI and Mistral have officially signed the code of conduct. This difference highlights the different strategies of tech giants in addressing EU AI regulations, as well as reflecting the dilemma of balancing compliance costs and innovation vitality in global AI governance.