On August 11, 2026, twelve of the country’s largest health systems announced they were joining forces with Aidoc, the clinical AI company best known for emergency imaging triage tools, to form what they are calling the Diagnostic AI Consortium. The systems involved, Advocate Health, Cedars-Sinai Health System, Hartford Healthcare, Houston Methodist, Mercy, Mount Sinai Health System, Northwell Health, Northwestern Medicine, Sutter Health, University of Florida Health, University Hospitals of Cleveland and WellSpan Health, together care for nearly 20 million patients a year. Rather than announcing a new algorithm, they are announcing a shared governance structure for the algorithms hospitals are already buying.
A Problem Bigger Than Any Single Tool
Hospitals have spent the past several years adopting diagnostic AI tools one at a time, often from different vendors, each cleared on its own validation data and deployed with its own local rules for how alerts reach clinicians. That piecemeal approach has left health systems with little shared language for comparing how well these tools actually perform once they leave a lab setting and enter a real emergency department. Aidoc CEO Elad Walach, describing the urgency behind the consortium in comments reported by HCInnovation Group, said the effort is about how the industry can move “quickly, but also responsibly,” adding that the real challenge with clinical AI “is not about the algorithm; it’s much more about the validation and diversity of workflows.” Walach also pointed to conditions on the ground, saying emergency departments are “in very dire straits these days,” a reference to the diagnostic errors and delays that overworked imaging and emergency staff are trying to manage.
What the Consortium Actually Plans to Do
According to the announcement, the group’s first job is to co-design AI-enabled diagnostic workflows meant to prioritize the most urgent cases, speed up interpretation, and get critical results to clinicians faster, while directly measuring the impact of those workflows on safety, quality and speed to diagnosis at each member site. The second and arguably more consequential piece is turning whatever works into shared implementation and governance practices that any hospital, inside the consortium or outside it, could eventually adopt. That is a notable departure from how most hospitals currently operate, where governance policies for AI tools are typically written system by system, with little cross-institutional comparison of outcomes.
The Technology Underneath
Aidoc is providing the technical backbone for the collaboration through two of its platforms: the Clinical AI Reasoning Engine, referred to internally as CARE, a foundation model built to reason across diagnostic findings, and aiOS, an enterprise operating system meant to let hospitals run and manage multiple AI tools across their imaging and diagnostic workflows from a single system rather than juggling separate vendor integrations for each one. The consortium’s structure suggests Aidoc is betting that hospitals want a platform for orchestrating diagnostic AI broadly, not just another point solution aimed at a single condition or scan type.
Why Now
The timing reflects a broader reckoning in radiology and emergency medicine, where AI adoption has outpaced consensus on how to evaluate it. Interpretation turnaround times for outpatient imaging have more than doubled over the past decade according to figures circulated in industry coverage, even as radiologists have left the field at markedly higher rates since 2020. Hospitals are under pressure to use AI to close that gap, but chief medical officers and radiology chairs have grown wary of adopting tools whose real-world performance is difficult to compare against each other, since each vendor tends to report results using its own validation studies and metrics. A shared consortium, in theory, gives member hospitals a way to pool outcome data across sites rather than each institution generating its own isolated evidence in a vacuum.
Skeptics Want to See Independence, Not Just Scale
Not everyone is convinced that a vendor-organized consortium is the right vehicle for independent evaluation. Because Aidoc is both the technology provider and a convening partner in the group, some health IT observers have raised the obvious question of whether governance standards developed in partnership with a single vendor can remain neutral enough to apply broadly across competing AI products from other companies. Walach has acknowledged the effort is “a multi-year journey” rather than a quick fix, and has said the consortium intends to publish its first findings within roughly a year, with governance meetings, site-specific pilot projects and shared learning built into the process from the outset rather than treated as an afterthought.
What Happens Next
The consortium says it expects to share its initial results sometime in 2027, which will be the first real test of whether pooling data and governance practices across twelve major health systems produces meaningfully different outcomes than each hospital continuing to evaluate diagnostic AI on its own. If the approach works, it could become a template other health systems and vendors look to as diagnostic AI keeps expanding into more corners of hospital workflows. If it stalls, or if member hospitals find the shared governance too generic to fit their own patient populations, it will reinforce how difficult it remains to build trust in AI tools whose behavior can vary significantly from one hospital’s patient mix to another’s.
