Google Pulls Dangerous Medical AI Overviews

12 Jan 2026 Earlier this month The Guardian reported that Google’s AI Overviews — short, automated summaries shown in search results — returned misleading and in some cases plainly false medical advice. The investigation highlighted examples experts called “really dangerous,” including an overview that told people with pancreatic cancer to avoid high‑fat foods (the opposite of some clinical guidance) and another that gave incorrect information about liver function tests, which could wrongly reassure people with serious liver disease. After those findings surfaced, the specific AI overviews in question appear to have been removed. As of this morning Google has disabled AI Overviews for queries such as “what is the normal range for liver blood tests?” A Google spokesperson, Davis Thompson, told The Verge that the company “invest[s] significantly in the quality of AI Overviews, particularly for topics like health,” and that an internal team of clinicians reviewed the examples shared with them and found that “in many instances, the information was not inaccurate and was also supported by high quality websites.” Thompson added Google works to make broad improvements where AI Overviews “miss some context” and takes action under its policies when appropriate. The Verge notes the update on January 11 added Google’s comment. The move is the latest controversy for a feature that has previously produced odd or harmful outputs (The Verge has catalogued past examples like telling people to “put glue on pizza” and recommending eating rocks) and that has faced multiple lawsuits. Source