On August 11, 2026, twelve of the largest U.S. health systems announced they are forming the Diagnostic AI Consortium, a joint effort with clinical AI company Aidoc to build shared standards for how diagnostic artificial intelligence is evaluated, deployed, and monitored across hospitals. The founding members, 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.
A Response to America’s Diagnostic Capacity Crisis
The consortium’s organizers frame the initiative as a response to what they call a widening gap between the volume of diagnostic imaging and lab work hospitals must process and the staff available to interpret it quickly. Rather than each health system independently piloting and validating diagnostic AI tools in isolation, often duplicating effort and reaching inconsistent conclusions about the same software, the twelve systems will co-design shared workflows, measure the AI’s real-world effect on diagnostic safety, quality, and speed, and publish common governance practices that other hospitals can adopt.
Why One Hospital Alone Isn’t Enough
Leonardo Kayat Bittencourt, vice chair of innovation at University Hospitals of Cleveland, one of the founding members, said no single center, however large or reputable, can capture the clinical and demographic diversity needed to prove that a diagnostic model performs safely everywhere it’s deployed. That concern echoes a broader problem regulators and researchers have flagged across cleared AI devices: a tool validated on one hospital’s patient population and scanner hardware does not automatically perform the same way at a rural clinic or a different health system’s imaging equipment. By testing the same AI models across twelve different environments simultaneously, the consortium aims to surface those performance gaps before they reach patients broadly.
Aidoc’s Role as Technical Backbone
Aidoc, founded in 2016, will supply the technical infrastructure underpinning the consortium’s work, including its Clinical AI Reasoning Engine, known as CARE, a foundation model built specifically for diagnostic reasoning, along with aiOS, an enterprise operating system designed to run multiple AI applications across a hospital’s imaging and clinical workflows. The company says its software already analyzes more than 60 million patient cases annually across roughly 2,000 hospitals worldwide. Aidoc raised $150 million in a Series E funding round in April 2026, pushing its total funding past $500 million, capital that has fueled an aggressive expansion beyond its original stroke and pulmonary embolism detection tools into a broader diagnostic reasoning platform.
Supporters See a Path to Trustworthy Scale
Hospital executives involved in the consortium argue that shared, transparent standards are the only realistic way to move diagnostic AI from scattered pilot programs into dependable, system-wide use. Their central bet is that AI can help identify urgent findings, a small pulmonary embolism, an early aneurysm, a subtle fracture, earlier than a busy radiologist working through a backlog might, and route those findings to the right clinical team faster. Doing that safely at scale, they argue, requires agreeing in advance on what counts as safe and effective, rather than each hospital defining success on its own after the fact.
Skeptics Question Vendor-Led Governance
Not everyone is convinced that a consortium built around a single vendor’s technology is the right model for setting industry-wide standards. Health policy researchers who have tracked the rapid growth of FDA-cleared AI devices, more than 1,000 as of early 2026, note that most of those tools were cleared without being tested against actual clinical outcomes, only against their ability to match a human reader’s interpretation. Critics argue that a consortium organized and technically powered by the vendor whose product is being evaluated risks producing standards that are easier for that vendor’s tools to meet than a truly independent, vendor-agnostic framework would allow. Others counter that involving twelve competing health systems, rather than Aidoc acting alone, provides a meaningful check against that risk.
What Comes Next
The consortium says it expects to share its first results in 2027, including data on whether the shared AI workflows measurably shorten the time between an abnormal scan being taken and a clinician acting on it. Member hospitals say they will also use the collaboration to build common training materials for radiologists and clinicians working alongside the AI, an area many hospitals have historically handled inconsistently or not at all. Whether the consortium’s governance model becomes a template that spreads to smaller, less resourced hospitals, or remains an initiative largely benefiting an already well-capitalized group of major systems, is likely to become clearer once the first outcome data arrives next year.
