Yale New Haven’s AI Competition Offers Model for Sourcing Solutions that Solve Real Problems

In an era where healthcare-focused AI startups abound, the most meaningful innovations may not emerge from Silicon Valley’s startup culture—but from inside the health systems themselves. That’s the premise behind the Health AI Championship, a novel competition led by Lee Schwamm, MD, SVP/Chief Digital Health Officer + Associate Dean, Digital Strategy and Transformation, Yale New Haven Health System & Yale School of Medicine.

Designed as a collaborative and transparent framework to stimulate responsible application development, the initiative serves as a replicable model for health systems seeking to harness AI while avoiding the pitfalls of vendor-led, siloed innovation.

A Model Rooted in Clinical Need and Fair Process

The Health AI Championship arose from Schwamm’s observation that most AI tools entering health systems are externally developed, often by technologists with limited understanding of healthcare’s core challenges. “The premise of this is threefold,” he said. “First, we want to drive solutions based on real clinical problems. Second, we want to reignite collaboration among health systems. And third, we want to create a fertile public-private environment where AI entrepreneurs want to work.”

Yale’s approach to the competition centers around problems defined by healthcare insiders. Each proposal submitted had to include a primary sponsor employed by one of Connecticut’s six major health systems. While collaborations with startups were allowed, applications needed to demonstrate a pre-existing, credible partnership—an intentional design to prevent unsolicited vendor-driven projects from dominating the process.

At the heart of the competition is a live pitch event featuring 12 finalists selected from 54 submissions. Winners stand to receive not only financial backing—a top prize of $100,000—but also the chance to validate their models using real-world data from Yale’s health system. The competition is funded through state-managed resources, including a grant originating from federal and public health innovation funds.



But beyond the prize money, the process itself is the true differentiator. Proposals were first ranked internally by each health system, which could submit up to five entries. Slots not used by one system were reallocated fairly across the others. Judges from each system then scored entries from their peer organizations—never their own—before the top-ranked proposals were handed to an independent judging panel for final review.

“Transparency was the core principle,” Schwamm said. “We had to build a process that participants from across systems could trust. And I believe we succeeded because fairness was embedded from the beginning.”

Driving Collaboration in a Competitive Landscape

While the health systems involved in the competition are natural competitors, the initiative created a rare moment of alignment. According to Schwamm, modern healthcare’s commercialization of ideas—particularly in AI—has discouraged open academic collaboration. “Data access has become a highly protected asset,” he said. “Instead of publishing research, people are forming companies. And that shift is understandable, but it comes at the cost of shared knowledge.”

Schwamm pointed to a broader trend where non-profit health systems, under intense financial pressure, are seeking revenue diversification through intellectual property. “That creates a walling off of ideas,” he noted, “where previously someone might have shared a promising model in a journal article, now they ask whether it can be monetized first.”

The competition aimed to reverse that trend by providing alternative incentives. “We’ve shown that if you structure an event transparently and provide meaningful validation opportunities, people are willing to participate and share,” Schwamm said. “Fifty-four submissions is a testament to that.”

One of the keys to success, he added,