Senate negotiators are discussing an AI oversight bill that could require AI
developers to demonstrate they have taken reasonable precautions to avoid
causing harm — a so‑called "duty of care" — and could empower the U.S. Commerce
secretary to demand such proof and to send government reviewers to test
products. What will qualify as "reasonable" measures is not yet defined; the
The proposal is under negotiation and its enactment into law remains uncertain.
