Manchin Seeks to Expand DOE Role in AI Research

Senate Energy and Natural Resources Committee Chair Joe Manchin (D-WV) expressed support for expanding the Department of Energy’s role in artificial intelligence research during a hearing last week. He said that DOE’s supercomputing resources and experience managing huge data sets give it a “natural leadership” role in AI, and he argued that the U.S. ought to build on efforts such as DOE’s Exascale Computing Project “rather than creating duplicate new programs at other agencies.”

Testifying before the committee, Deputy Energy Secretary David Turk reported that the department is proposing a major AI initiative called “Frontiers in Artificial Intelligence for Science, Security, and Technology,” which he said is based in part on a major report published this summer by six DOE national labs.

A lead author of the report, Argonne National Lab computer scientist Rick Stevens, also testified at the hearing, arguing that DOE is uniquely suited among federal agencies to lead development of AI systems for scientific and national security applications.

Committee Ranking Member John Barrasso (R-WY) expressed general support for DOE working to advance AI but focused his remarks on research security, specifically raising concerns about non-U.S. resident Chinese nationals working at DOE labs. Barrasso asserted that they are “beholden” to the Chinese government and could be coerced into stealing technology, such as through the government threatening their family members in China.

In response to a question from Barrasso on whether “the benefit of the work of the Chinese foreign nationals within our labs outweighs the documented risks,” Turk replied that DOE has many protections in place to prevent espionage, including a “risk matrix” that assigns additional restrictions to six sensitive technology areas, one of which is AI.

He also emphasized the important role that foreign-born scientists have long played in U.S. research.