The Promise of AI in Healthcare
Artificial intelligence has been heralded as one of the most transformative forces in modern healthcare. Proponents argue it can save lives by catching diagnoses earlier, automate tasks that have burned out an already strained workforce, and help address a growing physician shortage crisis. Yet the technology also carries serious risks — from overtrust that could endanger patients, to algorithmic biases that disproportionately harm underrepresented racial and ethnic communities.
Into this complex landscape stepped the Coalition for Health AI (CHAI) — a nonprofit with thousands of members united by a singular mission: to develop best practices for the responsible use of artificial intelligence in healthcare. At its founding, CHAI appeared to offer exactly what the industry needed: a credible, independent voice to help navigate the high-stakes world of health AI.
CHAI’s Rise and Early Ambitions
When CHAI launched as a nonprofit in March 2024, it drew immediate attention from Washington. Two senior Department of Health and Human Services (HHS) officials joined its board as nonvoting members. Major health systems — including Duke University Health System, Mayo Clinic, Stanford Health, and Johns Hopkins — pledged $1.25 million to the organization.
CHAI is led by Brian Anderson, M.D., a former family physician who helped found the organization in 2021 while serving as MITRE’s chief digital health physician. The group quickly attracted members spanning major health systems, community health centers, incumbent technology companies, startups, medical professional societies, and advocacy organizations.
The questions CHAI sought to answer were urgent: How can the FDA review a technology that changes daily? What safety standards does AI need before deployment in hospitals? How can rural America benefit equitably from AI innovation?
The Assurance Lab Initiative: A Vision That Collapsed
CHAI’s defining proposal was a nationwide network of AI assurance labs — independent testing environments where AI tools would be vetted for fairness, effectiveness, and safety before reaching hospitals and clinics. The concept, first outlined in a JAMA article in late 2023, drew on third-party assurance frameworks common in other regulated industries.
The Biden administration embraced the idea publicly. Former HHS Deputy Secretary Andrea Palm announced the administration’s support at the Health Datapalooza conference in September 2024. HHS even cited CHAI’s foundational JAMA publication in its Strategic AI Plan, published days before President Trump took office.
Yet behind the scenes, CHAI was struggling to deliver. By early 2025, the assurance lab concept had quietly collapsed. Anderson later acknowledged it was a misstep, saying the organization had incorrectly assumed that pre-procurement testing was the healthcare industry’s top priority. In reality, providers were more urgently concerned with post-deployment AI governance — monitoring models in real time to ensure they were safe, equitable, and performing as expected.
“Our initial hypothesis was that the pre-procurement use case was the one that would be most interesting to our doctors and nurses,” Anderson told Fierce Healthcare. “It wasn’t, as things turned out.”
The pivot to Assurance Resource Providers (ARPs) — vetted partner companies offering AI governance tools — was announced six months after the labs were abandoned, and with little public explanation. Industry executives interviewed by Fierce Healthcare criticized the lack of transparency and described CHAI as an organization that frequently adds new initiatives while failing to clarify the status of previous ones.
Political Fallout and Loss of Key Members
CHAI’s troubles deepened in the latter half of 2025, when its relationship with the Trump administration deteriorated sharply. In October 2025, HHS Deputy Secretary Jim O’Neill and FDA Commissioner Marty Makary published a scathing op-ed in the Washington Examiner, accusing CHAI of attempting to act as a “quasi-regulatory” body with inappropriate ties to the Biden administration.
Robert F. Kennedy Jr. amplified the criticism on social media, calling CHAI a “cartel.”
Shortly after, founding members Microsoft and Amazon departed the organization. Microsoft’s chief scientific officer stepped down from the board. The political controversy exposed CHAI’s vulnerability: an organization built on government partnership suddenly found itself on the wrong side of a new administration.
Despite the losses, Anderson reported that CHAI had not lost any provider organizations and had added new members, including the University of Texas, University of Virginia Health, and Emory.
What Comes Next for Health AI Governance
CHAI is now repositioning itself as a champion for healthcare providers and community health systems, having inked partnerships with the National Association of Community Health Clinics and the Joint Commission. It continues to run working groups on generative AI, prior authorization, and mental health chatbots, with roughly 1,000 unique individuals contributing to its workstreams.
Meanwhile, other organizations have stepped in to fill the gap left by CHAI’s abandoned assurance lab vision. URAC, the Consumer Technology Association, and the American Heart Association have all launched AI certification or assessment programs. A marketplace of third-party AI benchmarking is emerging — though it falls far short of the cohesive national network CHAI once envisioned.
For an industry still lacking clear federal rules on health AI, the stakes remain extraordinarily high. CHAI’s journey — from ambitious coalition to embattled nonprofit — reflects just how difficult it is to govern a technology moving faster than any single organization can manage.
