Stanford Computational Medicine Research Colloquium | When AI Writes: Fabricated References, Agentic Literature Synthesis, and Ambient AI
Zoom livestream: https://stanford.zoom.us/j/97887596012?pwd=bGVpbG1uL2ZBdEd2cHQzcVJMR3p6UT09
Webinar ID: 978 8759 6012 | Passcode: 420642
KEY INFO
When: Thursday, November 5, 2026, 12:00-1:00 p.m. Pacific
Where: Tapao Hall, 3180 Porter Drive, rooms B107 and B136, Palo Alto, CA 94304, or online via Zoom
Lunch will be served during talk | Q&A will follow
AFTER THE TALK
Recordings are posted on the ARISE podcast page within a week: https://www.arise-ai.org/podcast
SPEAKERS
Maxim (Max) Topaz, PhD, RN, MA, FAAN, FIAHSI, FACMI
Elizabeth Standish Gill Associate Professor of Nursing at Columbia University Medical Center & Columbia University Data Science Institute
Senior Research Scientist at VNS Health
IMIA Board Vice-President at MEDINFO2025 Taipei
Co-Director at Nursing and Artificial Intelligence Leadership (NAIL) Collaborative
Director of Services at Center for Community Engaged Data Science and Health Informatics
Maxim Topaz, PhD, RN, MA, FAAN, FIAHSI, FACMI, is a tenured Associate Professor at Columbia University School of Nursing, a Senior Research Scientist at VNS Health, and a health AI researcher. His career began as a combat medic and registered nurse, which shaped his commitment to improving patient care through technology. A Fulbright Fellow at the University of Pennsylvania and among the first nurse scientists to complete a postdoctoral fellowship at Harvard Medical School, he develops AI-based tools for patient prioritization and risk prediction, with a focus on home healthcare and unbiased AI. He is an international keynote speaker, past president of MEDINFO, co-chair of an NIH AI review section, and is recognized among Stanford's top 2% of scientists in health informatics.
THE TALK
"When AI Writes the Science: Fabricated References, Agentic Literature Synthesis, and Ambient AI"
Generative AI now drafts papers, synthesizes literature and listens in on clinical visits. This talk covers three lines of our work.
First, fabricated references: our audit of 2.5 million biomedical papers shows how citations to works that do not exist enter the published record and mostly stay there uncorrected.
Second, agentic literature synthesis: what AI agents can do when asked to read and summarize an entire research field, and where they fail.
Third, ambient AI in clinical documentation: what these tools capture, what they get wrong, and why clinicians still need to check the note.
The session closes with practical guardrails for researchers, journals and health systems.
HOST & MODERATOR
Hosted by the Stanford Division of Computational Medicine. Moderated by Jonathan H. Chen, MD, PhD, PI at Stanford ARISE.
KEEP IN TOUCH
Sign up for the free weekly Stanford Computational Medicine colloquia: https://forms.gle/LDMZP9RoVKXUbstC7
Explore Stanford HAILS: https://www.arise-ai.org/hails and HAILS Online: https://www.arise-ai.org/hails-online