U.S. Senate negotiators are discussing legislation that would require artificial intelligence companies to demonstrate that they are taking reasonable precautions to prevent their tools from causing harm, according to Reuters.
The proposal would give the U.S. commerce secretary the power to require developers to provide proof that they have taken reasonable steps to prevent harm, a standard described in the report as a “duty of care.” It would also authorize government auditors to test relevant products. Reuters said it could not immediately determine what would count as “reasonable,” and the legislation is still being negotiated.
Bill faces hurdles even if Congress approves it
Reuters said the measure would still face significant obstacles to becoming law even if it passes Congress. U.S. President Donald Trump said on Monday that existing authority is already enough to regulate and prosecute technology companies, signaling that he would not sign a bill setting new rules for artificial intelligence.
Senate leaders and committee members are involved
The proposal is being discussed by Senate Majority Leader John Thune, Senate Commerce Committee Chair Ted Cruz, and Democratic Senator Amy Klobuchar, who is part of the negotiations. Maria Cantwell, the senior Democrat on the Commerce Committee, is also involved in the talks.
Thune told reporters on Monday that Congress could put guardrails in place for more serious threats without hurting the United States’ lead in the AI race. In a statement, Klobuchar said she is continuing to push bipartisan legislation that would create government oversight for the biggest risks posed by AI models, including requirements for developers to work with government experts to validate and test models for safety.
Talks also cover a federal court route
Reuters reported on Friday that negotiators are also discussing a process under which the federal government, if it determines certain AI models are unsafe and wants to block their release, could take the matter to federal court. Companies would be able to challenge that decision in federal court.
The federal court portion of the measure would also block states from enforcing their own laws targeting specific risks tied to AI models.