U.S. Senate negotiators are considering bipartisan legislation that would require developers of the most advanced artificial intelligence systems to take greater responsibility for preventing potentially catastrophic risks.
The proposed legislation would establish a “duty of care” for AI developers, requiring companies to design and test advanced models with the aim of preventing serious threats. These could include AI being misused to assist in the development of biological or nuclear weapons or sophisticated cyberattacks.
Under the proposal, the U.S. government could potentially block the release of AI models considered unsafe. Companies would have the ability to challenge such government decisions in federal court, although lawmakers are still negotiating the details of the oversight framework.
The legislation could also restrict states from enforcing certain AI laws related to major model risks, creating a more unified federal approach to AI safety.
The discussions involve Senate Majority Leader John Thune, Commerce Committee Chairman Ted Cruz and Democratic Senator Amy Klobuchar. Senator Maria Cantwell has also pushed for advanced AI models to undergo testing by government scientists and experts.
With the November midterm elections approaching and only a limited number of congressional working weeks remaining, it is still uncertain whether lawmakers will reach an agreement and pass the legislation.
The negotiations come as concerns grow over the risks posed by increasingly capable AI systems and calls for stronger government oversight of advanced artificial intelligence.
Leave a comment