- The Washington Times - Monday, March 11, 2024

Call it the government’s first AI budget.

President Biden’s spending blueprint, sent to Congress on Monday, includes an infusion of money for artificial intelligence and a promise to tap AI’s potential while mitigating the risks posed by the emerging technology.

Mr. Biden asked Congress to approve $32 million for what he called the “AI talent surge,” an effort to hire computer and data scientists to explore AI and $70 million for agencies to hire chief AI officers who would be responsible for making sure the computer scientists are doing it safely.

The Energy Department said it is bringing AI to its fight against climate change. The Department of Veterans Affairs would get $10 million in AI money for medical and prosthetic research, and the National Nuclear Security Administration would receive $37 million to study AI misuse on chemical, biological or nuclear threats.

The Commerce Department would get $50 million to set up its AI Safety Institute, which is supposed to develop standards for using AI, including sorting out human-made and AI-generated content and working on privacy guardrails.

The budget envisions a total of $300 million in mandatory AI research funding governmentwide.

“Advances in AI are creating groundbreaking opportunities, while changing the nature of work and organizational management,” the White House said in the spending blueprint it submitted to Congress. “To benefit from the opportunities created by AI while mitigating its risks, the administration is committed to advancing its management of AI and significantly expanding AI talent in the federal government.”

The AI funding is a tiny but growing part of overall federal spending. Mr. Biden envisions reaching $7.3 trillion in fiscal 2025, which begins on Oct. 1.

It builds on last year’s budget, the first after the public release of ChatGPT and the accompanying AI craze.

Gregory Dawson, an Arizona State University professor studying AI spending with colleagues, said the budget still leaves the U.S. behind where it should be.

“Our concern remains the same as it was last year and the year before,” he said. “The U.S. has highly fragmented spending in AI, and this is in stark contrast to China, who is very focused in their spending. While this diffuse approach may eventually yield benefits, in the short term it continues to relegate us to second place behind China and losing more ground.”

A lot of what Mr. Biden plans to spend would be to carry out his October executive order on “safe, secure and trustworthy” use of AI by the government.

The budget also marks the beginning of an assessment of how AI might help evaluate how the government is handling big problems.

“Agencies will need to ask – and answer – evaluation questions such as: what is the impact of using AI on improved teacher productivity, student learning, and patient outcomes, as compared to current activities? Agencies must also evaluate the impacts of AI as it is deployed to improve targeting of government benefits and/or increase the reach of its programs,” the White House budget office said.

Many agencies seem to be in the earliest stages of feeling out AI, though most paid some lip service to the idea Monday.

The Justice Department said it was requesting $2.5 million “to capitalize on the extraordinary capabilities of artificial intelligence in a responsible and secure manner.”

Homeland Security Secretary Alejandro Mayorkas said he has earmarked $5 million for his department’s chief AI officer and a hiring “sprint” to enlist 50 AI experts in 2024.

“The new DHS ‘AI Corps’ will leverage this new technology across priority missions of the homeland security enterprise including efforts to counter fentanyl, combat child sexual exploitation and abuse, secure travel, fortify our critical infrastructure, enhance our cybersecurity, and deliver immigration services,” the department said.

A comptroller general’s report in December surveyed 20 agencies and found 1,200 current or planned uses of AI. NASA led the way with 390 uses, ranging from studying satellite data to managing the waiting room at a security ID badge office.

The Commerce Department ranked second at 285 uses, and the Energy Department was a distant third at 117 uses, followed by the Department of Health and Human Services and the State Department.

The report also highlighted the government’s struggles with AI, including some agencies’ delinquency in adopting safety regulations. The audit said the government’s hiring agencies at the Office of Personnel Management were still trying to develop a job classification for federal employees involved in AI work.

• Stephen Dinan can be reached at sdinan@washingtontimes.com.

Copyright © 2024 The Washington Times, LLC. Click here for reprint permission.

Please read our comment policy before commenting.

Click to Read More and View Comments

Click to Hide