Google’s search feature AI Overview cites YouTube more than any medical website when answering questions about health conditions, according to research, raising new questions about a tool viewed by 2 billion people every month.
The company has said that its AI summaries, which appear at the top of search results and use generative AI to answer users’ questions, are “reliable” And cite reputable medical sources like the Centers for Disease Control and Prevention and the Mayo Clinic.
However, a study that analyzed answers to over 50,000 health questions obtained using Google searches from Berlin found that the top cited source was YouTube. The video-sharing platform is the second most visited website in the world after Google, and is owned by Google.
Researchers at SE Ranking, a search engine optimization platform, found that YouTube accounted for 4.43% of all AI observation citations. He said no hospital network, government health portal, medical association or educational institution came close to that number.
“This matters because YouTube is not a medical publisher,” researchers wrote. “It’s a general-purpose video platform. Anyone can upload content there (e.g. board-certified physicians, hospital channels, but also wellness influencers, life coaches, and creators with no medical training).”
Google told the Guardian that the AI overview was designed to present high-quality content from reputable sources regardless of format, and that a variety of trusted health authorities and licensed medical professionals created the content on YouTube. It added that the study’s findings could not be extended to other regions because it was conducted in Germany using German-language questions.
The research comes after a Guardian investigation found that inaccurate and misleading health information in Google AI observation responses was putting people at risk of harm.
In a case that experts described as “alarming” and “alarming”, Google provided fake information about vital liver function tests, which could have misled people with severe liver disease into thinking they were healthy. The company later removed AI overviews from some but not all medical searches.
The SE Ranking study analyzed 50,807 health care-related signals and keywords to see which sources AI Overview relies on when generating answers.
They chose Germany because its healthcare system is strictly regulated with a mix of German and EU directives, standards and safety rules. “If AI systems rely heavily on non-medical or non-official sources, even in such an environment, it suggests the problem may extend beyond any one country.” he has written.
The researchers said AI observations appeared on more than 82% of health searches. When he looked at which sources AI Overviews trusted most for health-related answers, one result immediately stood out, he said. The most cited domain was YouTube with 20,621 citations out of a total of 465,823.
The next most cited source was ndr.deWith 14,158 citations (3.04%). The German public broadcaster produces news, documentaries and entertainment as well as health-related content. In third place was a medical reference site, msdmanuals.com With 9,711 citations (2.08%).
The fourth most cited source was Germany’s largest consumer health portal, Netdoktor.deWith 7,519 citations (1.61%). The fifth most cited source was a career forum for doctors, Praktischartz.deWith 7,145 citations (1.53%).
The researchers acknowledged the limitations of their study. It was conducted as a one-time snapshot in December 2025, using German-language questions, reflecting how users in Germany typically search for health information.
Results may vary according to time, region and wording of questions. However, despite those caveats, the findings still raise concerns.
Hannah van Colfschooten, a researcher specializing in AI, health and law at the University of Basel, who was not involved in the research, said: “This study provides empirical evidence that the risks to health posed by AI observations are structural, not real. This makes it difficult for Google to argue that misleading or harmful health outcomes are rare cases.
“Instead, the findings suggest that these risks are inherent in the way AI observations are designed. In particular, the heavy reliance on YouTube rather than public health officials or medical institutions suggests that visibility and popularity, rather than medical credibility, is the central driver for health knowledge.”
A Google spokesperson said: “The implication that AI Overview provides unreliable information is refuted by the report’s own data, which shows that the most cited domains in AI Overview are reputable websites. And from what we have seen in the published findings, AI Overview cites expert YouTube content from hospitals and clinics.”
Google said the study showed that 96% of the 25 most cited YouTube videos were from medical channels. However, the researchers cautioned that these videos represent less than 1% of all YouTube links cited by the AI overview on health.
“The majority of them (24 out of 25) come from medical-related channels such as hospitals, clinics and health organizations,” researchers wrote. “Additionally, 21 out of 25 videos clearly stated that the content was created by a licensed or trusted source.
“So at first glance this seems quite reassuring. But it is important to remember that these 25 videos are only a small part (less than 1% of all YouTube links cited in the AI overview). With the rest of the videos, the situation could be very different.”
