Trust at Scale: Safety, Transparency, and Feedback Loops for Clinical AI

Presentation Type: Panel
Presentation Time: 08:00 AM - 09:15 AM

Abstract Keywords: Analytical Artificial Intelligence: ML, Digital Pathology, Imaging AI, Predictive Analytics, Governance, Change Management, Outcomes Improvement and Equity, Generative AI in Clinical Workflow: Ambient Listening, Chart Summarization, Automated Response with LLM
Primary Track: Innovation through Industry, Public Health, Non-profit and Commercial Partnerships


Clinical AI has moved from pilot to production, yet safety practices have not kept pace with real‑world operational use. This panel focuses on the emerging frontier of operational AI safety—how health systems can detect, interpret, and respond to AI risk in ways clinicians understand and trust. Panelists will examine four practical questions: what safety signals AI products should expose (e.g., confidence, provenance, versioning); how model drift should be monitored in live clinical environments and what thresholds should trigger intervention; how AI errors should be reported through incident‑management pathways aligned with patient safety programs; and how model limitations can be communicated at the point of use without adding cognitive burden. Through case‑based discussion and audience engagement, the panel will synthesize lessons into a pragmatic Clinical AI Safety Starter Pack—a set of lightweight policies, monitoring elements, and transparency practices that any health system can adopt to improve safe, responsible AI deployment.

Moderator:
Joel Gordon, MD
UW Health, University of Wisconsin School of Medicine and Public Health

Speaker(s):
Aaron Grigg, MD
Grande Ronde Hospital

Keith Woeltje, MD, PhD, FAMIA
Froedtert ThedaCare Health

Shauna Overgaard, Ph.D.
Mayo Clinic

Author(s):
Keith Woeltje, MD, PhD, FAMIA - Froedtert ThedaCare Health; Aaron Grigg, MD - Grande Ronde Hospital; Shauna Overgaard, Ph.D. - Mayo Clinic; Joel Gordon, MD - UW Health, University of Wisconsin School of Medicine and Public Health;

---