Introduction: A New Era in Healthcare AI
Artificial intelligence is transforming how healthcare organizations deliver care. Epic, one of the world’s leading health technology companies, is pushing that transformation forward with the launch of Curiosity — a powerful new family of generative AI models. In a recent episode of AI Grand Rounds, Epic’s Seth Hain discussed how Curiosity builds on the company’s existing AI momentum. Together, these tools promise to change how clinicians predict outcomes and manage patient health.
What Is Curiosity?
Curiosity is a family of generative AI models developed by Epic. It represents a significant leap beyond traditional clinical decision support tools. Unlike rule-based systems, Curiosity learns from real-world patient data to generate predictions and insights.
Furthermore, Curiosity integrates seamlessly into Epic’s existing platform. This means clinicians do not need to switch tools or disrupt their current workflows. Instead, they gain access to richer, AI-powered insights right within their familiar environment.
How Curiosity Learns and Predicts
Training on Cosmos Data
Curiosity draws its intelligence from Cosmos, Epic’s expansive dataset. Cosmos includes more than 300 million deidentified patient records, making it one of the largest and most comprehensive medical datasets in the world.
The models learn from structured medical events. These include diagnoses, prescribed medications, and laboratory results. By analyzing patterns within these events, Curiosity predicts what may come next in a patient’s medical journey.
A Sequence-Based Approach
Rather than relying on isolated data points, Curiosity studies sequences of clinical events over time. As a result, it can identify subtle patterns that human clinicians might miss. This approach enables earlier interventions and more personalized care plans.
Key Use Cases for Clinicians
Seth Hain highlights several areas where Curiosity will make a measurable impact. These include:
Chronic Disease Management — Curiosity helps clinicians track disease progression and anticipate complications before they escalate.
Differential Diagnosis — The models assist physicians in narrowing down possible diagnoses by analyzing a patient’s full clinical history.
Hospital Operations — Beyond individual patient care, Curiosity supports administrators in managing capacity, staffing, and resource allocation more efficiently.
Each of these use cases demonstrates how AI can reduce cognitive burden on clinicians. Consequently, care teams spend more time on patients and less time on administrative tasks.
Curiosity and the Cosmos Research Community
Starting in February 2026, Curiosity becomes available to the Cosmos research community. Researchers gain access not only to the models themselves but also to advanced visualization tools. These tools allow researchers to explore projected patient trajectories and validate AI-generated predictions against real-world outcomes.
This open research approach accelerates innovation. Moreover, it ensures that Curiosity’s capabilities are rigorously tested before broad clinical adoption.
Epic’s Role in the Broader Healthcare Ecosystem
Connecting Devices, Diagnostics, and Therapeutics
Hain emphasizes that AI models alone do not deliver complete care. Devices, diagnostics, and therapeutics each play essential roles in the healthcare ecosystem. Epic’s network capabilities, particularly through Aura, connect these innovations directly into the clinician’s workflow.
Aura as a Clinical Bridge
Aura acts as a bridge between external innovations and point-of-care decisions. When a new diagnostic tool produces results, Aura integrates those results into the clinician’s dashboard. Therefore, the right information reaches the right person at the right time.
What This Means for the Future of Patient Care
The launch of Curiosity signals a broader shift in healthcare. AI is moving from narrow, task-specific tools to generative, predictive systems that understand the full arc of a patient’s health journey.
Additionally, this shift places greater responsibility on healthcare organizations. They must ensure that AI tools are used ethically and transparently. Epic addresses this by training Curiosity exclusively on deidentified data and by making the models available for peer-reviewed research first.
Patients, in turn, stand to benefit from earlier diagnoses, more targeted treatments, and better-coordinated care. Ultimately, this is what makes Curiosity more than a technical achievement — it is a step toward genuinely better health outcomes.
Conclusion
Epic’s Curiosity models represent a meaningful advancement in healthcare AI. They combine massive real-world data, advanced generative learning, and seamless clinical integration. As a result, they offer clinicians powerful new tools to predict, manage, and improve patient care. With the Cosmos research community now gaining access, the next chapter of healthcare AI has officially begun.
