Congress
Top Commerce Committee Democrat releases AI framework
Sen. Maria Cantwell, the top Democrat on the Commerce Committee with vast jurisdiction over artificial intelligence policy, outlined six AI safety principles Wednesday morning — tenets she believes should underpin any legislation to stave off the greatest threats posed by the rapidly-evolving technology.
“To manage risks from advanced AI systems we need clear safety standards, continuous testing, and reporting of serious failures,” the Washington Democrat said in a statement. “America can lead the world in AI by building systems that are not only more capable, but safer and more secure.”
Cantwell’s framework would encompass six priority areas: enforceable safety standards for frontier models, continuous testing, disclosure and oversight, public-private partnerships, protection for children and workers and adherence to U.S.-led global standards.
To directly address risks posed by the technology, her framework would have federal experts at the National Institute of Standards and Technology and other agencies establish clear federal rules to address harms, with specialized standards for open-source AI models to reduce risks of catastrophic misuse.
The senator is also recommending that the federal government provide developers with guidance on how to accelerate U.S. leadership on AI without sacrificing basic safety protocols.
On testing, Cantwell argues the government and independent experts should provide constant oversight and “continually scrutinize and stress test covered AI models,” particularly when sensitive information or secure facilities are involved.
She also says AI systems must have built-in safeguards to ensure they halt self-modification that could bypass safety controls. And she states that advanced models should not be released until they have undergone an independent, comprehensive audit — similar to outside auditors examining a company’s financial statements.
Cantwell, a former tech executive, has for years pursued similar priorities. But her new framework comes ahead of a midterm election that could return control of the Senate to Democrats, handing Cantwell back of the Commerce Committee gavel.
It also comes as lawmakers are under pressure to regulate the AI industry but are showing few signs that they will be able to reach consensus, across the aisle or even within their own party.
Senators had hoped for hearings, markups and briefings on AI legislation before leaving Washington until after the November elections, but little progress was made. Sen. Ted Cruz (R-Texas), the current Commerce Committee chair, has continued to say he intends to hold a markup on AI-related bills but has punted multiple times, saying senators have yet to reach agreement on any bipartisan measure.
Senate Majority Leader John Thune and Sen. Amy Klobuchar (D-Minn.) have been trying to cut a deal on legislation to prevent and contain catastrophic risk from runaway AI technology, but bill text has so far failed to materialize.
Cantwell’s approach laid out Wednesday morning would be stricter and more comprehensive than what Thune and Klobuchar are expected to ultimately produce, based on similar legislation they have offered in the past.
Her approach of “constant oversight and vigilance” would also serve as a contrast to that of GOP leaders, who have consistently called for a “light touch” regulatory regime.