Google has confirmed that it will sign the European Union’s voluntary Code of Practice for General‑Purpose AI, signaling its support for the EU’s AI regulatory framework despite expressing reservations.
Kent Walker, Google’s President of Global Affairs, conveyed that while Google sees value in the Code—particularly for supporting access to high‑quality, safe AI tools across Europe—it worries certain provisions might slow down innovation, expose trade secrets, or diverge from established copyright law, ultimately undermining Europe’s AI competitiveness.
What is the EU’s AI Code of Practice?
- It’s a voluntary framework, designed by 13 independent experts, to help companies align with obligations under the EU’s AI Act, especially Articles 53 and 55.
- It is organized into three chapters:
- Transparency—standardizes Model Documentation Forms for disclosing how AI models are trained and evaluated.
- Copyright—guidance for complying with EU copyright law in training data usage.
- Safety & Security—applies only to advanced “systemic‑risk” AI models and outlines technical and organizational safeguards.
- Published on July 10, 2025, with final approval by EU Member States and the European Commission expected later in 2025.
- Signing the Code is voluntary, but only signatories benefit from legal certainty under the AI Act. Non‑signatories must prove compliance through more onerous means.
Industry Split: Who’s Signing, Who’s Not?
- Google, OpenAI, Anthropic, and possibly Microsoft are signing or plan to sign.
- Meta has explicitly declined to sign, citing legal uncertainties, overreach beyond the AI Act, and concerns that the Code would curb innovation in European markets.
- Over 40 major European companies—including Airbus, ASML, BNP Paribas—have urged a two-year delay in enforcement, arguing that uncertainties around the Code could stifle competitiveness.
Why Google’s Signing Matters
- Strategic Compliance: Signing provides Google with a clear and harmonized path to compliance, minimizing the risk of being penalized under the AI Act.
- Credibility in EU Policy: It helps bolster trust with European stakeholders and underscores Google’s willingness to engage constructively—even while voicing feedback on problematic clauses.
- Competitive Signal: With Meta opting out, Google aligns with other major providers contributing to a unified approach in the European market.
What Are the Main Concerns?
- Trade secret exposure due to transparency requirements could expose sensitive business information.
- Potential deviations from EU copyright law —Google warns that overly broad or misaligned requirements might conflict with existing IP frameworks.
- Innovation slowdowns: strict approval processes or operational burdens may disadvantage European AI developers.
