Webinar #5 — Hybrid Human-AI Collective Intelligence in Medical Diagnostics

Photo by Antoni Shkraba Studio: https://www.pexels.com/photo/doctors-examining-an-x-ray-image-on-a-laptop-5214994/

4 November 2025

Addressing Core Weaknesses and Advancing the Capabilities of Today’s LLMs

In this webinar, hosted by the Human Diagnosis Project (Human Dx), we explored how collective intelligence, from human-only to human-AI hybrid collectives, is reshaping the way we approach intelligence now and into the future.

See also the video about the Human Dx platform that was not visible during the webinar, due to technical issues: https://www.humandx.org/product.

Speakers and talks

Dr. Eugene Chu, Kaiser Permanente

Eugene graduated from Yale with a BS in Biology, attended UC San Francisco (UCSF) for medical school, and completed residency in Otolaryngology and Head and Neck Surgery and a fellowship in rhinology and skull base surgery at Johns Hopkins Hospital. He completed a second fellowship in facial plastics/reconstructive surgery and anterior skull base surgery at UC Irvine. In addition to his work at Human Dx, he currently serves as Chief of Service of the Head and Neck Surgery Department at Kaiser Permanente and is the Downey Medical Center Area Research Chair. He is also a clinical assistant professor at UC Irvine and is involved in their facial plastic surgery fellowship training program. He has published over 20 peer-reviewed articles and 10 book chapters..

Past – Measuring and Modeling Human-only Collectives

Abstract: This presentation focuses on Human Dx’s prior work in scalably measuring clinician performance quantitatively and constructing physician collectives capable of outperforming even the best individual physicians.

Dr. Julian Berger

Julian Berger is an interdisciplinary researcher and postdoctoral fellow at the Max Planck Institute for Human Development in Berlin. He is affiliated with the Institute’s Center for Adaptive Rationality, where his work focuses on the intersection of human psychology, economics, and computer science to improve decision-making processes.

Present – Hybrid (Human-AI) Collectives

Abstract: In high-stakes medical decisions, deploying AI effectively while maintaining human oversight is a primary challenge. This talk explores simple and cost-effective heuristics for creating hybrid Human-AI intelligence, among others the use of Hybrid Confirmation Trees (HCT). Recent studies show how these approaches are powerful alternatives to both human-only and AI-only strategies, achieving comparable or superior diagnostic accuracy at a significantly reduced cost by leveraging the complementary strengths of human and machine cognition, specifically by capitalizing on their uncorrelated errors.

Jayanth Komarneni, Human Dx

Jayanth is the founder/CEO of Human Dx and has advised many healthcare stakeholders, including governments, foundations, payers, and providers. Previously, he worked at McKinsey and Bain before helping launch an alternative investment firm (billions in assets under management). Jayanth received his BS, BA, MS, and MBA from Penn and his MSc from Oxford. Jayanth participated in Y Combinator and is a Thouron Scholar, Luce Scholar, and Rhodes Scholarship finalist..

Future – What’s next?

Abstract: This presentation closes the webinar with a forward-facing perspective of the future of collectives in intelligence, not just in the medical domain, but in any expert system and beyond.