Leaked Anthropic Draft AI Regulatory Bill Is Unacceptable and Troubling (thezvi.substack.com)

A leaked draft of an AI regulatory bill proposed by Anthropic has raised significant concerns across the AI and machine learning community for its perceived deficiencies and troubling implications. The bill, which echoes elements of the earlier FRONTIER Act, is criticized for stronger preemption of state laws and reduced enforcement provisions. Critics argue that it prioritizes the interests of Anthropic over the safety of the public by making it harder for states to enforce regulations on AI development and deployment. With provisions that seemingly limit state powers to suspend potentially dangerous AI models, the bill has sparked outrage among advocates for robust AI safety regulations, who fear it may hinder necessary oversight. The significance of this development lies not only in its content but also in what it reveals about the current regulatory climate surrounding AI technology. Many experts believe that in light of accelerating AI capabilities and potential risks, regulatory frameworks should be more stringent, not weakened. Key shortcomings of the draft include inadequate measures for emergency shutdown capabilities and the lack of robust rulemaking for safety standards—elements that many argue are essential to mitigate catastrophic risks associated with AI. As the debate continues, stakeholders are urging a reconsideration of such proposals to ensure they genuinely address safety over corporate interests.