When you ask Google about your health, its AI often turns to YouTube

Google

• Google's AI health answers often cite YouTube over medical websites
• Experts warn YouTube lacks uniform medical standards for health advice
• Study finds AI Overviews may risk spreading inaccurate health information

Did our AI summary help?

When you ask Google about your health, its AI often turns to YouTube, and that is raising fresh concerns about how safe and reliable these AI-generated answers really are.

According to a report published by The Guardian, Google’s AI Overviews, the short summaries that now appear at the very top of search results, are more likely to cite YouTube than any medical website when responding to health-related questions. These AI summaries are seen by an estimated two billion users every month, giving them enormous influence over how people understand medical information.

The findings come from a study that analysed responses to more than 50,000 health queries captured through Google searches in Berlin. Researchers found that YouTube accounted for 4.43 percent of all citations used by AI Overviews. No hospital network, government health portal, medical association or academic institution came close to that level of visibility.

This is striking because Google has repeatedly said that its AI health summaries rely on reputable medical sources such as the Centers for Disease Control and Prevention and the Mayo Clinic. While those sources do appear, the study suggests they are often overshadowed by content from YouTube, a platform that hosts everything from licensed doctors to influencers and unverified health advice.

The researchers warned that YouTube is not a medical publisher and should not be treated like one. Unlike public health authorities or hospitals, it does not follow uniform medical standards, even though many videos appear authoritative and are designed to perform well in search results.

In one case described by experts as dangerous and alarming, Google’s AI Overview provided incorrect information about liver function tests. Doctors said the response could have led people with serious liver disease to wrongly assume their test results were normal. After criticism, Google removed AI Overviews for some medical searches, but the feature continues to appear for many others.

Hannah van Kolfschooten, a researcher specialising in AI, health and law at the University of Basel, said the findings show the issue goes beyond isolated mistakes. She said the risks posed by AI Overviews in health searches are structural and rooted in how the system is designed.

The study also raises questions about incentives. YouTube is owned by Google, and critics argue that the AI may be amplifying content from Google’s own platform instead of consistently directing users to trusted public health authorities.

As AI-generated answers become a routine part of online search, experts say the stakes are especially high for health information. When people turn to Google with medical questions, accuracy matters far more than popularity, and relying too heavily on YouTube could have real consequences.