According to Motion Intelligence Monitoring, Anthropic is driving gradual tightening of cutting-edge AI regulation in various U.S. states, advocating against a one-size-fits-all approach.
It has successively supported California's disclosure requirements, Illinois' annual third-party audit, and Massachusetts' stricter independent risk assessment. The latter requires evaluating AI systems for catastrophic risks such as assisting in developing bioweapons and allows for enforcement by the state attorney general.
On the other hand, OpenAI advocates for "reverse federalism." States would first adopt similar rules and then gradually form a nationally unified framework.
Despite not always being at odds, OpenAI has also backed Illinois' stringent bill. However, with Congress dragging its feet on action, state legislatures are becoming the main battleground where the two are vying for the direction of AI regulation.
