A 2024 paper and blog post unveiling Google’s healthcare AI model, Med-Gemini, mentioned “basilar ganglia” — a body part that doesn’t exist, The Verge reported.
After neurologist and researcher Bryan Moore, MD, discovered the mistake, Google fixed the blog post (though not the original paper), according to the Aug. 4 story. The company said it was a misspelling of “basal ganglia,” a part of the brain primarily responsible for motor control, but critics contend it points to the more serious issue of AI hallucination in healthcare.
“What you’re talking about is super dangerous,” Maulin Shah, MD, chief medical information officer of Renton, Wash.-based Providence, told the news outlet. “Two letters, but it’s a big deal.”
“Though the system did spot a missed pathology, it used an incorrect term to describe it (basilar instead of basal). That’s why we clarified in the blog post,” a Google spokesperson told The Verge. “We’re continually working to improve our models, rigorously examining an extensive range of performance attributes — see our training and deployment practices for a detailed view into our process.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
