A draft AI safety bill that Anthropic has been pitching to lawmakers leaked late last week, putting the company on the defensive. A company spokesperson told Semafor that the proposal is the legislative version of a safety framework Anthropic has promoted for months.
Safety advocates say the draft does not go far enough and would unfairly pre-empt state laws. An Anthropic lobbyist said the company plans to revise it to better reflect the pace of AI progress and its position on state laws. OpenAI’s Chris Lehane also argued that elected officials, rather than any one company, should lead policy decisions.
Why it matters
If the proposal keeps its state-law pre-emption provisions, states could have less room to set their own AI rules; the draft’s final terms are not yet clear.
