Anthropic published its Advanced AI Framework this June, laying out exactly what it thinks state and federal lawmakers should do about catastrophic risks from frontier AI. The company urged Congress not to block state-level AI laws unless and until it passes a rigorous federal regime , and even then, preemption should be narrow. States, Anthropic argues, should retain broad authority to legislate. For the crypto industry, this matters more than most realize.

The framework calls for mandatory safety testing, independent third-party evaluations, and public disclosure of risks for the most capable AI models. Anthropic explicitly endorses a state-by-state approach, joining OpenAI in a "reverse federalism" strategy , pushing model bills in states like California, New York, and Illinois to create a de facto national standard as federal action stalls. Illinois just passed SB 315, requiring frontier developers to publish safety frameworks and submit to third-party audits.

Why should crypto pay attention? Because every state-level AI bill that passes today becomes a template for tomorrow's AI-adjacent regulation. The same disclosure, audit, and testing infrastructure being built for AI models can and will be applied to AI-powered crypto protocols, algorithmic trading bots, smart contract auditors built on LLMs, and decentralized AI networks. The compliance architecture being debated in statehouses now will define the regulatory baseline that crypto projects will need to meet.

Moreover, the jurisdictional fragmentation that worries AI companies is an even bigger threat for crypto , a global industry that can't easily comply with a patchwork of 50 state regimes. If Congress eventually passes a federal AI law with preemption, crypto should fight to be included in that preemptive scope rather than left to state-by-state divergence. The regulatory ground being broken for AI today will almost certainly extend to crypto tomorrow. The industry should be watching , and engaging , while the blueprint is still being written.