Google to Sign EU’s AI Code of Practice, But Warns on Innovation Risks

Google announced it will sign the European Union's General Purpose AI Code of Practice. The company supports setting compliance standards under the upcoming AI Act but warns that certain framework elements could hinder innovation.

The search giant will join a voluntary framework created by 13 independent experts. The framework guides companies on requirements, including issuing summaries of content used to train general-purpose AI models and following EU copyright rules.

Kent Walker, Google's president of global affairs and chief legal officer, wrote in a blog post that the company wants the code to promote European access to secure AI tools. However, Walker cautioned that departures from EU copyright law, slower approvals, or requirements exposing trade secrets could reduce local model development and deployment, hurting competitiveness.

The European Commission unveiled the draft code of practice on July 10, 2025. Companies will be able to sign only after EU Member States and the Commission endorse the code, tentatively expected at the end of 2025. The voluntary framework provides legal clarity before the AI Act's phased rollout begins.

Key AI Act provisions apply on August 2, 2025, including transparency requirements for general-purpose AI models. Stricter rules for high-risk systems will follow in August 2026.

The voluntary code supplements the EU's Artificial Intelligence Act by offering guidance on transparency obligations and copyright compliance. Signatories receive a roadmap for documenting training data sources and obtaining permissions for content used in AI development.

In the past, Google estimated that prompt and widespread AI tool deployment under the code could boost Europe’s economy by around 8%, equal to €1.2–1.4 trillion by 2034.

Microsoft president Brad Smith told Reuters his company will likely sign the code. Meta Platforms declined, citing legal uncertainties for model developers. Meta's chief global affairs officer Joel Kaplan said in a LinkedIn post that the framework “goes far beyond the scope of the AI Act.”

The AI Act entered into force on August 1, 2024, becoming the first‑ever comprehensive legal framework on AI worldwide. The Regulation’s provisions will be introduced gradually over a six‑ to thirty‑six‑month period to ensure safety, transparency, bias mitigation and human oversight of high‑risk AI systems, especially in biometric applications and critical infrastructure management.