BIS Proposes New Reporting Rules for Leading AI Developers
New mandatory reporting rules proposed this week by the Bureau of Industry and Security could require developers of advanced artificial intelligence models and computing clusters to submit “detailed” information to the agency about their developmental activities, cybersecurity measures and more.
Sign up for a free preview to unlock the rest of this article
Export Compliance Daily combines U.S. export control news, foreign border import regulation and policy developments into a single daily information service that reliably informs its trade professional readers about important current issues affecting their operations.
The possible reporting requirements would help BIS and the Commerce Department better “understand the capabilities and security of our most advanced AI systems,” BIS Undersecretary Alan Estevez said. “It would build on BIS’s long history conducting defense industrial base surveys to inform the American government about emerging risks in the most important U.S. industries.”
The agency’s proposed rule, released Sept. 9, would implement an AI executive order signed by President Joe Biden last year (see 2310300029) by adding to the types of information BIS collects from U.S. companies in its industrial base surveys. The rule said BIS is specifically looking to collect data from businesses that are developing, have plans to develop or have the “computing hardware necessary to develop dual-use foundation models.”
Companies can integrate those models into military equipment, signal intelligence devices and cybersecurity software in ways that “enable those products to operate more effectively across a wider range of environments, to respond more effectively to unexpected signals, and to combat additional types of cyberattacks,” BIS said. “Given those potential capabilities, it is essential to the national defense that the defense industrial base” can integrate dual-use foundation models “to remain internationally competitive.”
To make sure the U.S. defense industrial base is prepared, BIS said the government needs more information about how U.S. companies are developing and using these models, information “about the characteristics” of models being developed and information about the cybersecurity measures companies are using to protect these models.
It’s also seeking information about the “safety and reliability” of these AI models, BIS said, especially because foreign adversaries could use them for activities that “threaten the national defense,” including by developing advanced weapons and technologies. The agency said it could request information from companies about the outcomes from their “red-teaming efforts,” which involve testing for dangerous capabilities of AI models, including their “ability to assist in cyberattacks or lower the barriers to entry for non-experts to develop chemical, biological, radiological, or nuclear weapons.”
“As AI is progressing rapidly, it holds both tremendous promise and risk,” Commerce Secretary Gina Raimondo said. “This proposed rule would help us keep pace with new developments in AI technology to bolster our national defense and safeguard our national security.”
Public comments on the proposals are due by Oct. 11.