OpenAI made big claims about transforming healthcare—experts are cautiously optimistic

Stargate is no longer just Gen X’s second-favorite Kurt Russell sci-fi movie (after The Thing, of course)—now it’s an initiative from some of the country’s largest tech companies that aims to develop the next generation of artificial intelligence (AI).

The Stargate AI infrastructure project, which President Trump announced in a press conference the day after taking office, is a collaboration between OpenAI and Oracle that has funding from Japanese company Softbank and United Arab Emirates-based investment firm MGX. The partners’ stated goal is to invest $500 billion over the next four years to build out the country’s AI infrastructure.

During the Jan. 21 press conference, OpenAI CEO Sam Altman and Oracle co-founder and Chief Technology Officer Larry Ellison claimed Stargate would revolutionize healthcare by doing everything from curing cancer and heart disease to making personalized cancer vaccines in just 48 hours.

Though developing that technology may be feasible, experts told Healthcare Brew that applying it is likely, for now, more “fi” than “sci.”

Not so simple. One statement Ellison made was that AI could be implemented into electronic health records (EHRs) to help doctors come up with “much better” care plans for patients.

The technology shows promise—early studies have shown AI can do as good of a job as a doctor at making accurate diagnoses, Pranav Rajpurkar, assistant professor of biomedical informatics at Harvard Medical School, told Healthcare Brew.

However, just because the technology works when researchers test it doesn’t mean it will be easy to implement across the healthcare system, according to Noémie Elhadad, chair of the biomedical informatics department at Columbia University’s medical school and an affiliate of the university’s Data Science Institute.

“There’s already fantastic improvements in how AI can help at the point of care. We can see them, but they’re not fully translated yet into care across multiple hospitals and populations,” Elhadad said. “That lack of applicability is really…a gigantic challenge.”

Part of that challenge is figuring out how to update healthcare payment systems to account for AI, Rajpurkar said.

“Who pays? The American healthcare system is complex. There are different stakeholders, each with different things that they value, and often the value to the patient isn’t necessarily aligned with the value of the system that’s being asked to pay,” he said.

Plus, hospitals will need to find the time and funding to train clinicians on how to use new AI technologies, Elhadad added.

“We could have the next 24- or 48-hour vaccine-of-all-cancer thing, and yet it would still not mean that it’s being used in practice by clinicians in the right way…just because clinicians don’t know how to use it,” she said.

Hoping for the best. It’s hard to say whether Stargate will be an overall positive or negative for healthcare, Elhadad said.

But, if it helps answer some of the big questions, like how to set up a payment system for AI, it could be a “net positive,” Rajpurkar said.

Elhadad added that, given the size of the investment, there’s reason for cautious optimism.

“I’m always excited when I hear that there is recognition, that there’s an opportunity…because we could really build these things,” she said. “I think the answer is to bring the right stakeholders to the room and make sure we’re going at it in a scientific way. But I am excited. I think it’s one of the most exciting times for AI in healthcare.”