The Biden administration spent much of this year laying the groundwork for how federal agencies will use artificial intelligence tools. Now the incoming Trump administration is charting its own course for how the federal government will use AI.
President-elect Donald Trump vowed on the campaign trail to repeal the Biden administration’s executive order on AI in government. In its place, the Republican National Committee supports “AI development rooted in free speech and human flourishing.”
Trump’s AI policy team is coming into focus through a slate of picks he announced earlier this month. Trump plans to appoint tech executive and investor David Sacks as his administration’s “AI & crypto czar.” Sacks and Elon Musk, one of Trump’s picks to head his Department of Government Efficiency, are both co-founders of PayPal.
Trump announced Sunday evening that Michael Kratsios, the former U.S. Chief Technology under the first term, will serve as the new director of the White House Office of Science and Technology Policy (OSTP) and an assistant to the president for science and technology.
Lynne Parker, former deputy U.S. CTO under the first Trump administration, will serve as the executive director of the Presidential Council of Advisors for Science and Technology (PCAST), and counselor to the OSTP director.
Former Microsoft executive Sriram Krishnan will serve as OSTP’s senior policy advisor for AI.
Trump signed several AI executive orders during his first term, and is likely to revisit the topic through additional executive action.
Federal agencies are seeing more and more use cases for artificial intelligence. Agencies have identified more than 1,700 potential uses of AI in their own operations. That’s about 500 more examples than agencies came up with a year ago.
The Departments of Health and Human Services and Veterans Affairs are exploring more AI use cases than other agencies.
Kurt DelBene, VA’s assistant secretary for information and technology and the department’s chief information officer, said VA’s AI use case inventory will continue to grow, and that, “it will just be natural that they are and will be part of things.”
“There are expectations, in terms of you have measured the effectiveness of that, that there aren’t any biases in the AI as well. So, expect us to ratch that up over time as well,” DelBene said last week at the FedFocus VA Digital Heath Care Summit.
Among its use cases, the VA is looking to pilot “ambient dictation,” or AI-powered note-taking that would take place during and after a veteran’s appointment with a VA clinician.
Nadia Smith, acting chief digital health officer at the Veterans Health Administration, said more than 200 vendors made AI pitches on Challenge.gov during the AI sprint, and that the VA plans to pilot ambient dictation in the “near future.”
“It is something that is helping to reduce administrative burden for our clinicians and our staff so that we can focus primarily on our veteran patients in the clinical setting,” Smith said.
OMB says government is generally seeing AI’s potential for internal agency support work health and medical work and helping agencies provide customer service to the public.
The Office of the Federal Chief Information Officer, in its yearend impact report, estimates the Biden administration spent $30 billion on federal AI use cases.
In March, OMB released its first governmentwide policy on how agencies should mitigate the risks of AI while harnessing its benefits. As part of OMB’s guidance, dozens of agencies have named chief AI officers, who are attending regular meetings of the Chief AI Officers Council to coordinate the development and management of AI across government
Agencies hired more than 250 AI experts into the federal government so far — more than halfway through the Biden administration’s goal of hiring 500 AI experts across the civilian federal government by the end of fiscal 2025.
Federal CIO Clare Martorana said in a recent interview that an “extraordinary amount of work” went into creating the AI use case inventory, but said more work is needed to accelerate AI across government. As for next steps, she said the Trump administration should consider a shared-services model to accelerate the use of AI across agencies.
“I don’t celebrate 1,700 use cases, candidly. It makes me sad, in some ways, that we have not solved some federal problems in a more universal manner — that we leave it to agencies to try to figure out how to do a chat tool in their own environment, without the benefit of learning from the people that have run faster,” Martorana said.
Martorana said her office’s AI team spent much of this year helping agencies understand governance, risk management and ethical use of AI. After doing all of that foundational work to get the federal government AI-ready, she said the next administration will make key decisions on how agencies use this emerging technology.
“I hope that that is some of the effort that is put in in 2025 is making sure that we understand, ‘Should there be a GovGPT? Should there be a model that we are using internal to government for pre-deliberative decision-making that benefits from all of the advantages of AI? Should there be procurements that we are doing that are universal? We have extraordinary computational capacity at our national labs. Are we using that to solve federal government problems?’” Martorana said. “Those are the questions that vex me, that I have tried to lean into and design some solutions for. And I hope that, again, the next team picks that up and really thinks about, ‘How do we use the incredible resources that we have across government to work like a team, to solve problems together and to drive the impact and deliver the experience that the American people expect of their government?’”
Among her recommendations for the next administration, Martorana urged the incoming team to not “make the same mistakes in AI that we have made in procuring IT.”
“We are doing everything from a blank piece of paper at every individual agency, like it’s the first time we’ve done anything, and that is not being the best stewards of the taxpayer dollars,” Martorana said. “I think that there needs to be more universal support for portfolios like this, as technology changes this rapidly, whether you call them shared services or vehicles that allow us to benefit from the work that’s being done at one fast moving, well-funded agency to benefit everyone.
Martorana said the AI use case inventory will give the incoming Trump administration a better understanding of where agencies see the most opportunity to rethink mission delivery.
“This organization has carried a huge load during this administration, and has really seized an incredible opportunity that future administrations are going to benefit from. Those are the reasons that it’s important to take a snapshot in time, take a picture, share that picture, so that the next team coming in understands the landscape, what they can build from, and then also what opportunities they have to clean the closets and rescind some things that have been on the books for a while, that are burdening teams with compliance. That isn’t driving outcomes. And that’s also something that we’ve attempted,” Martorana said.
Copyright
© 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.
