Sources: the US pushed back against EU's AI code of practice; the code is voluntary, still being finalized, and aims to help companies comply with the AI Act
Context & Ripple Effects
The code grew out of the EU’s first draft for general-purpose AI models, which paired copyright provisions with systemic-risk expectations. This report shows that the effort was not only a compliance exercise but also a point of transatlantic policy friction while its terms remained unsettled.
The later publication of the code’s documentation expectations made its compliance role more concrete, and the Commission subsequently maintained that it would not pause the AI Act rollout amid industry calls for delay [[a:887594|]].
First-order effects
- US objections add political pressure to the EU’s finalization of a voluntary instrument intended to guide AI Act compliance.
- AI developers considering the code face less certainty over its final terms and the degree of cross-border policy support behind them.
Second-order effects
- The dispute sharpens the choice for large AI providers between early alignment with the EU framework and waiting for clearer regulatory and diplomatic signals.
- Voluntary commitments can become a competitive differentiator: later participation by major providers demonstrated which companies were prepared to publicly align with the code.
Third-order effects
- If this pattern persists, AI compliance codes will increasingly function as arenas for competing jurisdictional influence, not merely technical implementation guides.
- The durable question is whether voluntary standards can bridge divergent national approaches or instead create separate compliance paths for global model providers.
The trend: AI governance is becoming a geopolitical contest over which jurisdictions shape the operating standards for globally distributed models.