Advertisement

National AI strategy resolution could pass during lame duck session

The authors would like to see an AI oversight framework, similar to the FITARA Scorecard, developed for agencies.
Dome of the U.S. Capitol Building, Congress, House, Senate, Washington, D.C.
(REUTERS / Erin Scott)

Reps. Will Hurd, R-Texas, and Robin Kelly, D-Ill., want to see their resolution to create a national artificial intelligence strategy passed this year, possibly during Congress’ upcoming “lame duck” session.

The House Committee on Science, Space and Technology is handling the resolution, and its authors are currently “twisting arms” for more co-sponsors, Hurd said during the launch of MITRE‘s Center for Data-Driven Policy on Thursday.

Quick passage of the resolution would provide the next couple of congresses with a framework for debating AI specifics, as the U.S. battles China to become the world leader in the emerging technology, Hurd said.

“The other idea is to take those 80-plus recommendations and turn those into individual bills,” Hurd said. “Can we narrow them as much as possible? Because that makes them easier to pass.”

Advertisement

Rather than establish a “Department of AI,” a national strategy should ensure every congressional committee has oversight plans and that every relevant agency has a role in organizing and securing data, procuring high-capacity compute and developing algorithms, Hurd said.

If the resolution is passed or individual laws are enacted, Hurd hopes Congress will oversee the adoption of federal AI much as it has with federal IT acquisition reform.

The outgoing congressman wasn’t in Congress when the Federal Information Technology Acquisition Reform Act (FITARA) was passed in late 2014 and said he was “suspicious” when the Government Accountability Office suggested a scorecard for measuring agencies’ compliance with IT reforms.

“But I can say that scorecard really did change behaviors, changed how the government operated, because we were keeping score,” Hurd said.

All agencies received passing FITARA grades for the first time in August, and Hurd and Kelly want to see similar oversight results with AI.

Advertisement

GAO is currently developing an oversight framework for algorithmic explainability, data quality and governance, and bias mitigation, said Chief Scientist Timothy Persons.

“I hope there’s some beautiful scorecard, like the FITARA Scorecard, that we can use,” Hurd said. “I don’t know what that is, but I think that’s the next evolution of this document and plan that Robin and I have been grateful to be able to work with.”

Latest Podcasts