NIST working on AI framework, ‘recognizing its power’

When discussing the creation of the Select Committee on Artificial Intelligence in May, Deputy U.S. Chief Technology Officer Michael Kratsios said it would be premature to begin considering regulating AI this early. But some federal leaders and lawmakers think now is precisely the time to consider the implications of the new technology.

In fact, Walter Copan, under secretary of Commerce and director of the National Institute of Standards and Technology, said during a July 11 conversation with Politico about AI that NIST is already working on developing a trustworthy, understandable framework around AI.

“AI is indeed with us,” he said. “We are recognizing its power.”

Copan said the federal government has a series of roles to play regarding AI. It needs to be involved in setting the research agenda, it needs to weigh in on the workforce implications, and it needs to help coordinate with the private sector and academia to ensure an open and transparent process in developing standards.

That’s why NIST is already working with stakeholders to test parameters to demystify the uses of AI. Copan said the goal is to fully leverage the human-AI interface. It needs to be seen as a trustworthy ally to human activity in areas related to the federal space, like agriculture, health care and mobility.

And Copan isn’t the only one thinking along these lines. Rep. John Delaney (D-Md.), co-chair of the AI Caucus, also pushed back on the idea that it’s too early to be planning for AI.

Sputnik-level response needed to artificial intelligence

“I do think that [the administration’s] notion that this isn’t already upon us or their view that this isn’t already affecting our lives in very profound ways, or to have the view that decisions aren’t being made today that will have very significant effects on how this technology will play out is in fact shortsighted, because I think the toothpaste is already out of the tube a little bit as it relates to how artificial intelligence is affecting our lives,” Delaney said. “And I do think, very similar to the way the United States responded to Sputnik, kind of a national strategy around building the space program, and aligning the government sector and the private sector towards a set of common goals, I think that would be very very useful right now at this moment in time.”

Advertisement

Delaney said some countries like China and France already have national AI strategies, and the U.S. needs to remain competitive to protect its citizens. It doesn’t have to be a regulatory approach right away. The government, he said, needs to be active in confirming that the programming embedded in AI is consistent with our values. It also needs to start reaching out to allied countries and begin the conversation around international standards and developing a common language, because AI will have profound effects on international trade and will cross political boundaries.

Rep. Will Hurd (R-Texas), concerned about competition from China in the AI space, suggested some concrete steps the federal government could start taking immediately to begin moving in the right direction. He said the government needs to direct federal research money to AI, unlock its data for private industry to begin taking advantage of, and begin adopting and using AI tools in government as test cases.

Delaney did, however, make it very clear that the need for this has nothing to do with the fear that we may suddenly find ourselves in a situation reminiscent of the Terminator movies.

“People often go right away to the doomsday scenarios with these new technologies,” he said. “And I think what they don’t realize is that there’s a million steps between now and then where there’s an opportunity for human intervention. And what will make that human intervention most successful is if there’s real transparency.”