President Trump Monday proposed cutting federal research spending—except in key areas including artificial intelligence and quantum technologies.

Trump’s budget for the fiscal year beginning October 1 proposes spending $142.2 billion in research and development, 9 percent less than in the current year. The White House says its proposal is 6 percent more than it requested last year.

The budget request is something of a gambler’s approach to funding American innovation, betting big in select areas. “I find it disappointing and concerning that funding for basic research is down,” says Martijn Rasser, a senior fellow at the Center for a New American Security, a policy think tank in Washington, DC. “We just don’t know where the next breakthroughs will come from.”

The budget goes all-in on AI and quantum, proposing to double funding across the departments including the National Science Foundation, the National Institutes of Health, the Department of Energy, Darpa, and the DoD’s Joint AI Center.

At the same time, the president proposed cutting research spending in nearly every arm of the government, including by $424 million at the National Science Foundation, $4.7 billion at the Defense Department, and $3.2 billion at the Department of Energy.

Rasser says it’s good to see proposed funding for AI and quantum increase, since both are critical emerging areas. He recently coauthored a report that called for substantially more government funding for AI, noting that the technology could transform industries in the way that software has.

But Rasser worries that cutting overall funding for fundamental research will undermine progress and growth because innovations can emerge from many fields. “It’s just not a good trajectory,” he says.

The president’s budget is a proposal and is likely to be altered by Congress. But the plan signals an approach to scientific research that prioritizes areas that are seen as crucial to international competition and military advantage. AI and quantum technologies have received substantial investment from America’s big geopolitical rival, China, and both are being rushed into military service by Beijing.

AI has emerged as a powerful technology, with big tech companies investing billions into its development and harnessing it to build everything from self-driving cars to voice assistants.

Quantum computing and communications technologies, built on the peculiar properties of quantum physics, are far less proven, though they have the potential to deliver equally big payoffs someday. A quantum version of the internet, for instance, should guarantee perfect security, while quantum computers could crack now-unsolvable problems with relative ease.

During Trump’s time in office, AI and quantum technologies have become increasing areas of focus. The president launched the American AI Initiative in February 2019, outlining a national strategy for AI leadership. And Trump signed the National Quantum Initiative Act into law in December 2018, increasing funding for quantum computing and communications.

Today, AI researchers find themselves in hot demand and, increasingly, able to secure ready government funding. But some are wary of the White House plan.

“It looks like a mixed blessing,” says Subbarao Kambhampati, an AI professor at Arizona State University and a past president of the Association for the Advancement of Artificial Intelligence.

Kambhampati notes that the funding allocated for AI seems geared towards driving fundamental work in AI, which will help push new advances in the field. “I am happy to see funding for basic research,” he says.

But he warns that cutting funding for other areas of research may not only undermine progress in other fields but also have a knock-on effect on AI, because progress has often drawn from other areas. Studies of animal brains can inspire new designs for artificial neural networks, for exampl
e. “Neuroscience will be really relevant to AI in over the longer term,” he says.

Kelvin Droegemeier, director of the White House Office of Science and Technology Policy, defended the proposed drop in research funding in a call with reporters. “President Trump has made very clear that investing in early stage basic research is extremely important and this budget does that,” Droegemeier said. “I think it's a very responsible act on President Trump's part to make sure that we prioritize what are high importance activities are especially relative to great power competition.”

Dario Gil, director of IBM research and a member of the President’s Council of Advisors on Science and Technology, argues that government investment in AI will boost scientific discovery in general, thereby helping advance research in other areas. “The purpose is to accelerate discovery,” he says. “That is going to permeate every sector of the economy and national defense.”

Gil also suggests that, like the internet before it, quantum computers and communications networks may have unforeseen payoffs. “We need to see a significant acceleration of the investments, and this is a very good step in the direction,” he says.

He might be right, but it’s tricky to pick the winners in research and innovation. As Rasser says, “With basic research, you never know what’s going to pan out.”