Senate aides are debating a bill that would impose a duty of care on AI developers, potentially giving the government power to block unsafe models.
Senate aides are weighing a draft bill that would obligate artificial‑intelligence developers to take a “duty of care” in addressing known, high‑impact risks, a move that could give Congress the authority to halt the deployment of models deemed unsafe.
Proposed duty of care framework
The legislation, still in early negotiation, would require AI firms to conduct risk assessments, implement mitigation strategies, and report findings to the Federal Trade Commission. If a model is judged to pose a “major risk” to public safety, national security, or civil liberties, the bill would allow regulators to order its suspension or modification.
Key concerns driving the proposal
Lawmakers cite incidents ranging from disinformation campaigns to biased hiring tools as evidence that existing voluntary safeguards are insufficient. The bill aims to create enforceable standards that mirror those applied to pharmaceuticals and financial services, where companies must prove safety before market entry.
Critics argue that the language could be overly broad, potentially stifling innovation and giving the government excessive discretion over emerging technologies. Industry groups have warned that vague definitions of “known major risks” might lead to inconsistent enforcement.
Legislative path and timeline
The proposal is expected to move to the Senate Judiciary Committee later this month, where it will be reviewed alongside bipartisan input from technology experts. If approved, the measure could be incorporated into a broader AI oversight package slated for a vote before the end of the year.
- Risk assessment and mitigation plans
- Regular reporting to the FTC
- Authority for the FTC to suspend unsafe models
- Penalties for non‑compliance
Stakeholders from academia, civil‑rights groups, and the tech sector have been invited to submit comments, ensuring the final language balances safety with the need for continued AI research and development.
“We need clear, enforceable standards that protect the public without choking innovation,” said a senior Senate aide involved in the drafting process.
The outcome of these negotiations will shape how the United States regulates advanced AI systems, setting a precedent that could influence global policy discussions on technology safety.