Professor warns of dangerous failure modes in medical AI
Gadeau described label‑confusion failures in medical imaging—where spurious cues such as rulers can drive diagnostics—and cited IBM Watson's struggles to generalize beyond elite datasets as a caution for clinical AI adoption.
The full story
At the FLCC webinar, Dave Gadeau explained how seemingly small dataset artifacts can produce hazardous clinical errors: when images used to train models include rulers or other consistent marks, the model can learn to use those marks as predictors instead of underlying pathology. "If you don't have a ruler on your chest at that time, AI is not gonna identify it as a tumor," he said, describing published study examples and mitigation efforts.
Sign up to keep reading.
Your free newsletter brings you what your city council, county, and school board discussed and decided.
Keep reading for free
Create a free account to finish this story and see the video, transcript and key details.
Continue with GoogleContinue with emailAlready have an account? Log in