Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
At a time when the administration of US President Donald Trump is trying to slash federal spending on science, Darío Gil is in an enviable position. As under-secretary for science for the US Department of Energy (DoE), he presides over programmes that the administration prizes: those advancing artificial intelligence and quantum science.
Last week, his agency announced its goal of building the world’s first ‘fault tolerant’ quantum computer for solving scientific problems by 2028, in response to Trump’s executive order on quantum innovation. Also on Gil’s to-do list: boost researchers’ confidence in AI through the administration’s US$600-million Genesis mission, which launched last November.
As part of Genesis, the DoE has been tasked with developing an overarching AI platform, imbued with a variety of models, that can connect and query scientific instruments, supercomputers and data sets at the country’s 17 national laboratories. The goal is to use this architecture to tackle scientific challenges in collaboration with researchers at universities and private companies.
Gil acknowledges that enthusiasm for the initiative has not been uniform, but says this is to be expected: academics “are educated to be sceptical”.
Still, some are eager — or, they’ve gotten the memo that this is where they can find funding. The DoE’s first call for Genesis proposals in March garnered a record 5,000 applications, which is 2.5 times more than the agency has ever received for a funding call, Gil says. “It’s a lot of enthusiasm.” Next month, the agency will announce a limited pool of winners, to which it will collectively distribute $293 million in funding.
Addressing scepticism
Gil moved from Spain to the United States as a secondary-school exchange student. After earning a PhD in electrical engineering and computer science at the Massachusetts Institute of Technology in Cambridge, he took a position at computer giant IBM in Yorktown Heights, New York, where he eventually climbed the ranks to research director and senior vice-president. He joined the DoE last September.
It’s clear from chatting with Gil that he has a fundamental love of science. So how does he respond to fears that basic science in the United States is being cut in favour of AI research? At the DoE, he says, the pinch that researchers might be feeling comes from a long-term trend at the agency of increased spending on maintaining user facilities (which host specialised instruments for the whole scientific community) and building new ones. This is a dynamic “that is independent, right now, of Genesis”, he says. This spending has been at the expense of research grants, and it could be rebalanced, he adds, but by making cuts to user facilities.
In total, the DoE’s Office of Science budget, per research area, is up compared with last year, Gil says. The office’s contribution to Genesis, spread across two rounds of investment so far, is $520 million, with some of the mission’s cash coming out of research budgets for basic-science programmes, such as nuclear and high-energy physics. But Gil insists that because Genesis projects will span a variety of scientific disciplines, funding will still be allotted to those areas — projects will just have to include AI. Between disciplines, “I haven’t moved one dollar,” he says. “All I am asking is that if you’re doing biology, can you please explore carefully and thoughtfully the implications of this computing revolution for biology?” he says.
Another challenge for Gil is addressing scientists’ worries about AI safety. The Trump administration has mostly taken a hands-off approach to AI regulation. Earlier this month, however, faced with the rapidly growing capabilities of frontier models, it effectively forced Anthropic to take the Fable 5 model offline just days after its release, by ordering the company to suspend access to foreign nationals. Gil doesn’t directly answer questions about regulation, instead he advocates for AI firms and the government to work together to mitigate threats. “Deep scientific collaboration at the frontier of AI is a huge mechanism to actually also harness, protect it and shape its future — as opposed to being in a position where you’re just reacting,” he says.
Enjoying our latest content?
Log in or create an account to continue
Access the most recent journalism from Nature's award-winning team
Explore the latest features & opinion covering groundbreaking research