Artificial intelligence is rapidly evolving from systems that generate answers to autonomous agents capable of searching, reasoning, and supporting complex biomedical decisions. Yet realizing the potential of agentic AI in medicine requires more than increasingly capable models: it requires systems that are trustworthy, evidence-grounded, and rigorously evaluated in real-world settings.
Dr. Zhiyong Lu,
NIH
In this talk, I will present our recent efforts to develop trustworthy AI agents across the spectrum of biomedical research and clinical decision support, from gene set analysis (Nature Methods, 2025) and deep evidence synthesis (Nature Machine Intelligence, 2026) to clinical trial matching. Together, these examples illustrate how AI agents can move beyond generating fluent responses toward systematically finding, evaluating, and synthesizing biomedical evidence to support scientific discovery and clinical decision-making.
Dr. Zhiyong Lu is a Senior Investigator at the National Library of Medicine where he directs biomedical AI and machine learning research for biomedicine. In addition, Dr. Lu is Adjunct Professor of Computer Science at the University of Illinois Urbana-Champaign (UIUC). Dr. Lu is a highly cited researcher with over 450 publications (H-index: 99), and his AI research has been deployed in real-world systems like PubMed and LitCovid, benefiting millions each day. Dr. Lu’s research has been frequently featured in major news outlets and recognized with numerous awards, including the NIH Director’s Award, Clinical Center CEO Award, and NLM Regents Award. Dr. Lu has been elected to the American College of Medical Informatics (ACMI) and the International Academy of Health Sciences Informatics (IAHSI).