A leaked AI bill raises a question about who can say stop
Zvi Mowshowitz objects to a purported Anthropic bill draft. He argues it may curb state AI safety rules without giving federal officials enough power to audit or halt a risky system. He also quotes other views. The text is a leaked draft, not an enacted law or an independent legal review.
What the source reports
Zvi Mowshowitz argues that a purported Anthropic bill draft would replace some state AI safety powers with a weaker federal plan. The question is not just which level of government makes the rules. It is whether the new plan gives reviewers access to evidence and a real way to stop a dangerous system. His essay also quotes analysts who differ on parts of the plan. This is commentary about a leaked draft. It is not enacted law, and the digest does not offer an independent legal reading. A fair test would compare the draft text with current state rules and the federal powers it would add.
Original source
- Title
- Leaked Anthropic Draft AI Regulatory Bill Is Unacceptable and Troubling
- Author
- Zvi Mowshowitz
- Publication
- Don't Worry About the Vase
- Date
- Saturday, October 17, 2026