How Anthropic is pursuing a state-by-state push for ever-tougher AI safety laws, in contrast with OpenAI's “reverse federalism” strategy for common state rules
SAN FRANCISCO — Artificial intelligence giant Anthropic is pursuing a strategy of one-upmanship that encourages states …
Context & Ripple Effects
The companies’ state-level approaches have diverged amid a longer fight over whether Washington should preempt state AI rules. Anthropic previously lobbied Congress against a proposed ban on state AI regulation, while OpenAI has promoted a strategy of using state legislation to establish a broadly shared baseline.
The split is also visible in liability policy: Anthropic opposed an Illinois proposal backed by OpenAI that would limit lab liability even in severe-harm cases. That makes the current advocacy contest more than a disagreement over venue; it is a disagreement over how stringent AI accountability rules should be.
First-order effects
- State policymakers and advocacy coalitions now face competing industry-backed models: Anthropic’s push for stronger state-by-state safeguards versus OpenAI’s effort to align states around common rules.
- Anthropic and OpenAI will be judged by lawmakers not only as AI developers but as opposing policy sponsors, particularly on the trade-off between liability exposure and protections for AI labs.
Second-order effects
- AI legislation may become less uniform across states as competing proposals use different thresholds for safety duties, liability, and enforcement, raising compliance complexity for developers operating nationally.
- Other AI firms, enterprise customers, and insurers may be pressed to take positions on whether to support a common baseline or stricter state-specific obligations, rather than treating AI regulation as a single federal-policy question.
Third-order effects
- If states continue to set consequential AI rules without federal preemption, the US could develop a durable state-led AI governance system in which influential jurisdictions establish practical national standards through market access.
- The companies’ disagreement suggests that “AI safety” will remain an active competitive and legal-policy battleground: the key structural question is whether state experimentation converges on a common floor or entrenches a patchwork of liability and safety regimes.
The trend: This is one data point in the shift from abstract federal AI-policy debate toward companies competing to shape state-level rules that can become de facto national standards.