Republican Senators See AI as a Shortcut to Shrinking the Government

Republican senators are eager to implement artificial intelligence tools into federal government operations as a way to cut down on spending, even if they’re not sure what it would look like in practice.

Federal agencies have rolled out a number of AI initiatives in recent months, and President Donald Trump has made it clear he wants the U.S. to lead in the sector. The Republican senators who recently appropriated billions toward AI initiatives in Trump’s massive domestic policy law have plenty of ideas on how federal agencies could use AI. But they have yet to grapple with some of the broader issues the technology is running up against in other industries.

“I see AI as being an agent of speed and coordination,” Sen. Mike Rounds, a member of the Senate Committee on Armed Services, told NOTUS. “You want AI to give you a speedier process through which you can save money.”

Rounds, like most Republican senators, voted to give $1.5 billion in their recent reconciliation package to the Department of Defense to invest in AI tools for shipbuilding, cybersecurity and the development of AI-powered autonomous weapons.

Like other senators, he had a general idea of how he would like to see it applied.

“When it comes to defense and strategy overall, you want AI to look at multiple options in terms of what your adversary might be doing,” Rounds said. “AI can be the agent to speed up any of those processes if you train them accurately.”

But many senators who have high hopes for AI have only vague familiarity of what its day-to-day applications look like in the agencies they hope to use it in.

In the Republican reconciliation package, border authorities received $6.1 billion to, among other purposes, contract AI tools to detect narcotics. In a conversation with NOTUS, Sen. Rick Scott said he would “have to hear details” from border authorities about the implementation of AI, but he had high hopes for how it could make federal agencies more efficient.

“AI is an excellent predictive technology,” Scott said. “I mean, when I was in business, I tried to use predictive technology for quite a while.”

“What AI is good for is that it gives us potential opportunities to do our jobs better,” Scott added. “If we’re going to spend money on any technology, we ought to be saying. ‘How do we provide a better service at a lower cost?’’

Several agencies have publicized efforts to modernize some of the old software that much of the federal digital infrastructure runs on, and to make government databases compatible with AI in an effort to ease the workload of a reduced federal workforce.

In May, the Department of Energy announced the acquisition of supercomputers to pursue AI research. In June, Department of the Interior officials suggested they would use AI to go through probate court backlogs. Last month, the Environment Protection Agency said it would start using AI to analyze large databases. The Department of Education also announced it was exploring ways to incorporate AI into the education system. And the DOD announced multiple million-dollar contracts with AI developers to create software tools for autonomous weapons and manufacturing.

Even so, AI adoption has been a nuisance for some agencies like the Food and Drug Administration, which recently deployed an AI chatbot intended to aid scientists who work at the agency to speed up their lengthy government research and drug approval process. Last month, CNN reported that the agency’s chatbot was not connected to the internet, limiting its research capabilities, and has provided researchers with inaccurate citations.

Sen. Dan Sullivan told NOTUS that keeping up with advancements in the technology and integrating it into the federal bureaucracy “is a challenge.” Sullivan argued that massive agencies like the Pentagon run the risk of adopting redundant and inefficient software that creates additional “bureaucratic roadblocks.”

However, he believes AI could become crucial to keeping the federal government from falling behind in its key functions.

“Let me give you an example of shipbuilding. We have a really decrepit system,” Sullivan said, adding that he just walked out of a hearing touching on this very subject. “So if there’s something from an automation perspective that can make our processes more efficient using some of the AI technology, then absolutely we need to use it.”

In just a few years, AI chatbots like ChatGPT have amassed hundreds of millions of weekly users. But despite the mass adoption of AI among students, lawyers, the technology sector and many lawmakers on Capitol Hill, the technology still makes many mistakes — some of them suspected to have made it into government reports — and there have been reports of more than a few alarming user interactions.

And Congress has been unable to create any guardrails for the development or deployment of AI technology both at the consumer and industry levels. While GOP lawmakers recently tried, and failed, to block state-level regulations (many said they preferred a federal standard), congressional Republicans haven’t brought forward any comprehensive AI regulation proposals yet.

Still, AI’s popularity is one of the reasons why some GOP lawmakers think the government should become a key player in the industry.

“AI is already being incorporated in virtually every aspect of our lives, and the speed of adoption will increase exponentially. I think that’s inevitable,” Sen. Ted Cruz, one of the AI industry’s main allies in the Senate, told NOTUS.

Trump has touted a comprehensive plan to accelerate AI development and adoption in the U.S. by creating new rules for government agencies to only contract AI developers that design their models that “pursue objective truth rather than social engineering agendas,” according to the report, and “eliminate references to misinformation, Diversity, Equity, and Inclusion, and climate change.”

For some Republicans, AI promises a future that challenges many people’s long-standing view that the federal government moves at too sluggish a pace.

“Think about that for a second,” Sen. Bernie Moreno told NOTUS as he explained how passport issuing could be sped up by AI. “You’ll someday have to explain it to your grandkids, ‘I swear! There was a booklet of paper and it was bundled with a seal and it had my picture on it.”

“We’re not using experimental technology, we’re using technology that’s been fully commercialized,” Moreno added.