TL;DR
Following a Guardian investigation that flagged misleading and false medical answers, Google has disabled some AI-generated overviews for certain health queries. Experts called the examples alarming and dangerous; Google declined to comment on the specific removals.
What happened
A recent investigation by The Guardian found that Google’s AI-generated overviews returned misleading or plainly false information for some medical queries. Two examples highlighted by experts included advice for people with pancreatic cancer to avoid high-fat foods — which specialists said was the opposite of appropriate guidance and could raise mortality risk — and incorrect details about liver function tests that might lead people with serious liver disease to believe they were healthy. In response, the AI overviews for at least some medical questions — for example, queries such as “what is the normal range for liver blood tests?” — appear to have been disabled. Google declined to comment on the specific removals. The episode adds to a string of controversies around the feature, which has previously produced odd or dangerous suggestions and has been mentioned in multiple lawsuits.
Why it matters
- AI-generated summaries can reach broad audiences and, when incorrect, may cause direct harm to people seeking medical guidance.
- Search providers are increasingly expected to manage the accuracy of automated answers, raising questions about oversight and safety.
- Removal of results signals internal or public-pressure remediation, but also underscores the difficulty of guaranteeing correct clinical guidance from general AI models.
- Ongoing legal and reputational risk for companies that surface automated health information without reliable safeguards.
Key facts
- The Guardian published an investigation showing misleading and false medical answers in Google’s AI overviews.
- Experts described some of the example answers as “alarming” and “dangerous.”
- One cited instance: the AI advised people with pancreatic cancer to avoid high-fat foods; experts said this advice was the opposite of appropriate guidance.
- Another instance involved incorrect information about liver function tests that could mislead people about serious liver disease.
- AI overviews for questions such as “what is the normal range for liver blood tests?” have been disabled.
- Google declined to comment on the specific removal of those AI overviews.
- The feature has previously produced bizarre or harmful suggestions in other contexts, and has been the subject of multiple lawsuits.
- The Verge reported the removals and summarized the Guardian’s investigation.
What to watch next
- Whether Google will restore, revise, or permanently disable AI overviews for medical queries — not confirmed in the source.
- Any public explanation, postmortem, or technical fixes from Google addressing how the errors occurred — not confirmed in the source.
- Potential regulatory or legal actions related to AI-generated medical content following these disclosures — not confirmed in the source.
Quick glossary
- AI overview: A short, automatically generated summary provided by an artificial intelligence system in response to a user query.
- Liver function tests: Blood tests that measure enzymes and proteins to help evaluate how well the liver is working; used in clinical assessment.
- Pancreatic cancer: A condition in which malignant cells form in tissues of the pancreas; clinical guidance on nutrition and treatment varies by situation.
- Misinformation: False or inaccurate information that can mislead readers, whether shared intentionally or not.
Reader FAQ
What did Google remove?
AI-generated overviews for at least some medical queries, including questions like “what is the normal range for liver blood tests?”
Why were they removed?
Google declined to comment on the specific removal; the change followed a Guardian investigation that exposed misleading and false medical answers.
Were the AI answers actually dangerous?
Experts cited in the investigation described some of the answers as “alarming” and “dangerous,” including one case involving pancreatic cancer guidance and another involving liver tests.
Is this feature facing legal action?
The source says the feature has been the subject of multiple lawsuits.
Will Google fix the system?
Not confirmed in the source.
NEWS AI SCIENCE Google pulls AI overviews for some medical searches Experts described the results as “alarming” and “dangerous.” by Terrence O'Brien Jan 11, 2026, 7:19 PM UTC 0 0…
Sources
- Google pulls AI overviews for some medical searches
- 'Dangerous and alarming': Google removes some of its AI …
- Google removes AI Overviews for certain medical queries
- Google removes multiple AI health summaries 'after results …
Related posts
- Google’s AI Inbox for Gmail offers task-focused view, early hands-on
- Google unveils Universal Commerce Protocol to enable AI-agent shopping
- Industry Insiders Launch Poison Fountain to Corrupt AI Training Data