PALO ALTO, Calif. – The National Institutes of Health is seeking comment from industry on a new artificial intelligence strategy for the agency.
Principal Deputy Director of the NIH Matthew Memoli, M.D., announced the RFI at the Coalition for Health AI’s semi-annual meeting at Stanford Medicine on Thursday. The forthcoming plan adds to the pro-AI stance of the administration and to explicit actions to center the use of the technology with the Department of Health and Human Services (HHS).
The RFI seeks comment on the use of AI in biomedical discovery, public health and clinical decision support. And the agency wants feedback on how the NIH can best collaborate with other government agencies, non-governmental advocacy organizations and industry on AI development and testing.
“The NIH is building the first NIH AI strategic plan, which will guide AI research funding initiatives in healthcare for years to come across the entire agency,” Memoli said. “We're going to be issuing an RFI to inform our AI strategy. The RFI will mark the start of an ongoing public-private partnership to shape the NIH approach to AI,” he said.
Memoli said that he and NIH Director Jay Bhattacharya, a career Stanford clinician, are excited about the use of AI in the healthcare industry and for use in the federal government. He emphasized President Donald Trump’s commitment to U.S. leadership in AI with the signing of his January executive order.
Internally, the NIH is working to develop a strategy to “create a unified central AI structure to build synergy across program silos, improve transparency and accelerate R&D translation … to patient benefit,” Memoli said.
It plans to develop a one-year action plan and a longer-term plan for how the agency can use and support AI in the healthcare ecosystem, Memoli said.
NIH seeks comment on how it should develop its strategic architecture like “data readiness, trust, translation and workforce” that can work as the foundation of the plan. It also asks stakeholders to comment on actions it can take to transition to using semi-autonomous agents and the use of AI for hypothesis generation, reproducible studies and continuous learning.
In addition to identifying foundational principles for the NIH’s strategic AI plan, the NIH is also looking to collaborate with other government agencies and industry to simplify AI resource creation and use of tools.
One section of the RFI says NIH will engage with the Food and Drug Administration and the Department of Veterans Affairs for regulatory science research for AI, testbeds and evaluating clinical AI tools. It also seeks comment on metrics it can use to evaluate AI pilots at the agency.
The RFI emphasizes consensus-based standards, transparency and trust in the use of AI. It asks about models for shared governance, licensing and maintaining jointly developed AI assets.
An attendee of the CHAI member meeting that works with the VA told Memoli the department was also seeking to engage with CHAI, like the NIH has. “We may be able to build centralized relationships so each agency doesn't have to set up their own individual relationships,” Memoli said.
The troves of data that will be needed to continue to fuel AI development could come from research done at the NIH, he said. Memoli also said that the NIH should conduct research on how to implement AI in healthcare.
“We really are going to have to do the research through NIH to understand how healthcare and AI can be used, because we don't want to just start instituting AI without understanding what we're doing and showing that we can demonstrate the benefits of it. … so that we can understand how, where and when to use it in what setting, in what situation, to get the maximal benefit.”
The administration is just starting to get to work on policies that affect the healthcare technology industry. This week, HHS leaders have cemented their excitement about the use of AI to aid research, ease data access and improve the function of HHS.
In addition to the creation of an AI strategic plan at the NIH, Health Secretary Robert F. Kennedy Jr. and CMS Administrator Mehmet Oz, M.D., talked about the use of AI to streamline issues in health data interoperability at a meeting with stakeholders on June 3.
“[The meeting] was revealing of the administration's priorities in three ways,” Morgan Reed, CEO of the App Association, said. “The fact that you had an event with the Secretary and the head of CMS, both spending time there says that this is an administration priority. The second is that they are doubling down on the idea that AI is there not to be an interlocutor with the patient, but to help with all of the kind of cruft, all of the thick cotton that gets in the way of doctors serving patients or patients getting the information that they need. And the third overarching priority is a focus on improving patient care and outcomes. It was clear patients are at the center of all they are going to do."
William Charnetski, executive vice president for health system solutions and government affairs at PointClickCare, said the administration intends to centralize the use of AI to solve long-standing issues.
"From what I observed and heard, this administration is committed to digital health transformation – with artificial intelligence at its core," Charnetski said in a statement to Fierce Healthcare. “Among the most compelling themes is the recognition that collaboration with the private sector, open communication, and interoperable data exchange must be treated as foundational, not optional. Access to accurate, timely health information is not merely a technical goal but a life-extending imperative. Equally important is the acknowledgment that the responsible use of AI must be underpinned by strong guardrails and broad access to high-quality data. These priorities resonate deeply with my own views and underscore the transformative potential of AI to unify data across fragmented care environments.”