A bipartisan group of congressman on Wednesday unveiled legislation that would require federal agencies and their artificial intelligence vendors to adopt best practices for handling the risks posed by AI, as the U.S. government slowly moves toward regulating the technology.
The proposed bill, sponsored by Democrats Ted Lieu and Don Beyer alongside Republicans Zach Nunn and Marcus Molinaro, is modest in scope but has a chance of becoming law since a Senate version was introduced last November by Republican Jerry Moran and Democrat Mark Warner.
If approved, the bill would require federal agencies to adopt AI guidelines unveiled by the Commerce Department last year.
It would also require the Commerce Department to develop specific standards for AI suppliers to the U.S. government and call on the Federal Procurement Policy chief to develop language to require those suppliers "to provide appropriate access to data, models, and parameters... to enable sufficient test and evaluation," the bill says.
Generative AI, which can create text, photos and videos in response to open-ended prompts, in recent months has spurred excitement as well as fears it could make some jobs obsolete and upend elections by making it difficult to distinguish between fact and false information. In extreme examples, there is concern AI could allow bad actors to access critical infrastructure.
The United States has taken tentative steps toward regulating AI but Europe is much more advanced.
U.S President Joe Biden signed an executive order last October to make AI safer, requiring developers of AI systems that pose risks to U.S. national security, the economy, public health or safety to share the results of safety tests with the U.S. government, before they are released to the public.
