REGULATION · US
Anthropic proposes AI safety framework requiring judicial approval for model blocking
Anthropic put forward a policy proposal that would require federal agencies to obtain a judge's approval before blocking an AI model release, even if the model poses high-risk harms. The bar: evidence that 50+ deaths could result.
WHY IT MATTERS
Signals how frontier labs are shaping AI governance discourse: framing safety guardrails as requiring judicial due process rather than executive prerogative, potentially raising compliance burden and litigation risk for regulators.
Source: TechCrunch · 2026-10-20