Radiology has become the single biggest foothold for artificial intelligence in American medicine, with the overwhelming majority of FDA-cleared AI medical devices aimed at reading scans. Now a dozen U.S. health systems have banded together in a new collaborative designed to accelerate how diagnostic imaging AI gets built, validated, and deployed, with Aidoc, one of the field’s best-known imaging AI companies, playing a central role.
Why Radiology Became AI’s Biggest Beachhead
Imaging has been the natural starting point for medical AI because the data involved, X-rays, CT scans, and MRIs, is already digital, standardized, and available in enormous volume, exactly the conditions that make machine learning models easiest to train and validate. The U.S. Food and Drug Administration’s list of authorized AI and machine-learning medical devices now runs into the thousands, and radiology alone accounts for roughly three-quarters of all of them, reflecting both the technical head start and the clear clinical need for faster scan interpretation.
What the New Collaborative Is Trying to Fix
Despite that head start, health systems have struggled to move diagnostic AI from a single pilot department into consistent, hospital-wide use. Each health system historically has had to independently validate imaging AI tools against its own patient population and workflows, a slow and duplicative process that the new collaborative is designed to streamline by letting member hospitals share validation data, implementation lessons, and governance frameworks rather than each reinventing the process from scratch.
Aidoc’s Role and Perspective
Aidoc CEO Elad Walach has described the consortium’s purpose around using AI to achieve faster, more accurate diagnosis across member hospitals, framing the collaborative less as a product rollout and more as a coordinated effort among health systems to professionalize how they adopt diagnostic AI tools generally. Aidoc’s own platform flags abnormalities on scans, such as signs of stroke or pulmonary embolism, for urgent radiologist review, and the company has positioned itself as an early convener of this kind of cross-institutional cooperation.
The Case for Shared Standards Over Solo Pilots
Hospital-by-hospital validation of AI tools has been criticized within the industry as inefficient and inconsistent, producing a patchwork where a tool proven effective at one health system must essentially prove itself all over again at the next. A shared collaborative approach could, in principle, establish common benchmarks and implementation playbooks that smaller or less resourced hospitals, which often lack the data science staff to run their own validation studies, could adopt more easily than if they had to negotiate directly with each AI vendor on their own.
What Skeptics Want to See Before Celebrating
Collaboratives convened partly by a vendor with a commercial stake in the outcome, as Aidoc has here, invite a fair question about whose interests the group ultimately serves. Independent radiologists and health policy researchers have generally welcomed efforts to standardize AI validation but caution that shared frameworks need transparent, vendor-neutral governance to avoid simply making it easier for a handful of established AI companies to lock in their market position under the banner of standardization, rather than genuinely improving patient outcomes across all vendors.
What Happens Next for the Twelve Health Systems
The collaborative’s success will ultimately be measured not by its membership roster but by whether participating hospitals can demonstrably speed up adoption timelines and show measurable diagnostic improvements compared to systems going it alone. If the model works, it could become a template other specialties facing similar AI validation bottlenecks, from pathology to cardiology, look to replicate. For now, the initiative remains in its early stages, with the real test still ahead as member hospitals begin sharing data and implementation experience in practice rather than principle.
Photo: Charlss GonzHu / PEXELS via Pexels
