Twelve of America's largest health systems have formed an unprecedented alliance to solve a quiet but urgent crisis: patients are waiting longer than ever for diagnostic imaging results, and the radiologists who interpret those scans are leaving the field faster than new ones arrive. The Diagnostic AI Consortium, launched by Aidoc in partnership with systems including Cedars-Sinai, Mount Sinai, Northwestern Medicine, and University Hospitals Cleveland, collectively serves nearly 20 million patients annually and is designed to prove that artificial intelligence can safely accelerate diagnosis without sacrificing accuracy.

The timing reflects a genuine healthcare bottleneck. Between 2014 and 2023, interpretation turnaround times for outpatient imaging more than doubled, according to the Harvey L. Neiman Health Policy Institute. At the same time, radiologists have been leaving the field at a 50% higher rate since 2020, and researchers project the shortage will persist through 2055 without deliberate action as an aging population drives imaging demand faster than new radiologists enter the workforce.

What Makes This AI Partnership Different From Past Healthcare Tech Initiatives?

The Consortium represents a fundamental shift in how healthcare organizations approach AI deployment. Rather than each hospital system testing diagnostic AI in isolation, the 12 members are committing to shared standards for evaluating, validating, and governing AI tools before they scale. This collaborative approach addresses a critical gap: no single health system, however large, can capture the diversity of imaging, disease patterns, and clinical contexts needed to build AI models that work reliably across different patient populations, scanner types, and geographic regions.

"For a long time, the industry treated speed and safety as opposing forces in AI: Silicon Valley's 'move fast and break things' against medicine's 'first, do no harm.' This Consortium is built on the belief that it can be done the right way, guided by two principles: iteratively and together," said Elad Walach, CEO and co-founder of Aidoc.

Elad Walach, CEO and co-founder of Aidoc

The Consortium will operate through three core functions. First, members will co-design AI-enabled diagnostic workflows intended to prioritize urgent cases, accelerate image interpretation, and deliver critical results faster to improve patient flow in real-world clinical settings. Second, they will measure the impact of these workflows on safety, quality, and speed to diagnosis across all member sites. Third, they will convert what works into shared implementation and governance practices that any health system can adopt, inside the Consortium or beyond it.

How Will the Consortium Actually Test and Deploy Diagnostic AI?

The technical backbone comes from Aidoc's CARE clinical AI foundation model and aiOS enterprise AI operating system, which manage deployment, workflow integration, and post-deployment monitoring. With AI already running in nearly 2,000 hospitals and analyzing 60 million patient cases annually, Aidoc brings operational experience at a scale no single institution can match. The Consortium expects to share its initial results in 2027.

The partnership reflects a broader recognition across healthcare that diagnostic AI requires a different approach than consumer-facing AI. Diagnostic AI must detect subtle clinical signals across imaging, pathology, laboratory data, and medical records, helping physicians recognize disease earlier and act faster. In medicine, where every missed or delayed diagnosis can change a patient's outcome, that requires AI purpose-built for clinical decision-making, rigorously validated in clinical practice, FDA-cleared, and designed to improve measurable patient outcomes.

"Joining the Diagnostic AI consortium reflects our belief that this next generation of AI in radiology will only reach its full potential through shared expertise and accountability across institutions and the industry. No single center, however large or reputable, can capture the diversity of imaging, disease, and clinical context needed to build and deploy models that generalize safely and at the scale the world needs," said Leonardo Kayat Bittencourt, MD, Vice Chair of Innovation at University Hospitals.

Leonardo Kayat Bittencourt, MD, Vice Chair of Innovation at University Hospitals

Steps to Implement Diagnostic AI Safely in Healthcare Settings

• Establish Shared Governance Standards: Multiple health systems must agree on common evaluation criteria, validation protocols, and performance monitoring practices before deploying AI tools, ensuring consistency across institutions and patient populations.
• Monitor AI Performance for Drift and Bias: Continuously track how AI models perform across different patient demographics, scanner types, and clinical sites to detect performance degradation or systematic errors that could harm specific populations.
• Integrate AI Into Clinical Workflows: Design AI tools to fit naturally into existing diagnostic processes, prioritizing urgent cases and accelerating interpretation without disrupting physician decision-making or adding administrative burden.
• Measure Real-World Impact: Collect data on safety, quality, and speed improvements across member sites to generate evidence that AI deployment actually improves patient outcomes, not just processing speed.
• Share Implementation Lessons Broadly: Document and publish what works so other health systems can adopt proven practices, accelerating responsible AI adoption across the industry.

The Consortium's formation also signals a maturation in how healthcare organizations think about AI governance. Every day, physicians must synthesize expanding volumes of information across imaging, pathology, laboratory data, and medical records, often under mounting operational pressure. AI performance must be continuously monitored for drift and bias across patient populations, sites, and scanners, a discipline still rare in AI broadly. The Consortium's members are holding themselves to shared standards precisely because they recognize that scaling diagnostic AI responsibly requires discipline that goes beyond what any single vendor or institution can enforce alone.

The initiative arrives as health systems move beyond isolated AI applications toward enterprise-wide diagnostic AI. Members share a vision of what that shift makes possible: a health system where a critical finding starts moving through the care pathway soon after a scan is complete, where the sickest patients are correctly prioritized, and where a diagnostic workup that stretches across days is completed in hours. The Consortium's findings are intended to augment and advance initiatives already underway across the industry, from Aidoc and its customers to regulators, medical societies, and health systems building their own programs.