Key Players and the EU AI Code of Practice
The European Union's new
AI Code of Practice is drawing mixed responses from major tech companies.
Microsoft is likely to sign the code, positioning itself alongside Europe's push for responsible AI development and compliance with the upcoming AI Act. In contrast,
Meta has publicly declined to adhere to the voluntary guidelines, marking a sharp policy divide among industry leaders[3].
Overview of the EU Code of Practice
- The AI Code of Practice is a voluntary framework developed to help companies align with the EU's AI Act, which becomes enforceable starting August 2, 2025[1][2].
- Key requirements include:
- Respecting copyright and transparency in model training.
- Responding to opt-out requests from creators whose work may be used as training data.
- Regularly publishing documentation describing AI features and systemic risks[1][3].
- By signing, companies can demonstrate their commitment to compliance, potentially benefiting from lower administrative burdens and increased legal certainty[1][2].
Microsoft’s Approach
Microsoft is expected to be among the first major industry players to commit to the EU's framework. The company emphasizes its ongoing collaboration with European regulators to implement the AI Act effectively and responsibly[4][5]. By signing the code, Microsoft will signal its intent to prioritize safety, transparency, and legal compliance as the regulation landscape evolves.
Meta’s Rejection and Implications
Meta has openly rejected the EU's guidelines, arguing the measures may be too restrictive or burdensome[3]. Mark Zuckerberg's company faces the prospect of greater regulatory scrutiny in the EU, as those who opt out of the code must find alternative, equally robust ways to prove compliance with the AI Act.
Risks and Penalties for Non-Compliance
- Companies that do not adhere to the AI Code of Practice may be subject to increased examination by EU regulators[3].
- Violations of the AI Act can result in strict financial penalties:
- Up to 7% of a company's annual sales for some breaches.
- Up to 3% for developers of advanced AI models[3].
Looking Forward
The divide between
Microsoft and
Meta on voluntarily signing the AI Code of Practice reflects an evolving regulatory climate for AI in Europe. While the code remains voluntary, its alignment with the region’s AI Act means that the choices tech giants make now could heavily influence their regulatory and reputational standing in the years to come.
