Imagine a tool trusted by billions that might be leading us toward inaccuracies—this is precisely the concern when it comes to Google’s AI Overviews providing health information. And here’s where it gets controversial: recent findings reveal that YouTube, the video-sharing giant owned by Google, is cited more frequently than established medical websites when answering health-related questions through Google’s AI summaries. This raises critical questions about the reliability of the information being presented, especially considering how popular and accessible YouTube is for the average user.
Google maintains that its AI-generated overviews are based on reputable sources, including distinguished institutions like the Centers for Disease Control and Prevention (CDC) and the Mayo Clinic. These summaries appear at the top of search results, designed to give quick and trustworthy answers. However, a detailed investigation into over 50,000 health queries originating from Berlin’s Google searches paints a different picture—one where YouTube accounts for a significant share of the citations.
Researchers from SE Ranking, a platform specializing in search engine optimization, discovered that YouTube was responsible for about 4.43% of all citations in these AI summaries. That might seem modest at a glance, but it becomes alarming when compared to official health organizations or academic sources, which accounted for far less. In fact, no hospital network, government health portal, or renowned medical organization approached that figure.
The key issue here is the nature of YouTube itself. Unlike medical journals or official health websites, YouTube is a general-purpose video platform open for anyone to upload content—ranging from licensed physicians and hospital channels to wellness influencers, life coaches, and individuals with no medical credentials at all. This diversity poses a risk; it becomes unthinkable to rely entirely on a platform that hosts both credible medical professionals and untrained content creators.
Google responded to these concerns, emphasizing that AI Overviews aim to highlight high-quality, reputable sources regardless of the content format. They pointed out that many health authorities and licensed professionals produce content on YouTube. However, the fact that this study was confined to German-language queries in Germany highlights a limitation: regional differences could lead to varied results elsewhere.
This investigation is particularly troubling because it follows another alarming report from The Guardian, which revealed that Google’s AI summaries sometimes deliver misleading or even false health advice, potentially endangering users. One especially concerning example involved misinformation about liver function tests, where users might have been led to believe they were healthy despite serious underlying conditions. Google did take corrective action by removing some of these AI summaries, but not all, illustrating a broader risk.
To deepen our understanding, researchers examined nearly 51,000 keywords related to health, focusing on Germany due to its strict healthcare regulations—making the findings even more noteworthy. When analyzing the sources AI Overviews relied upon, the most cited domain was YouTube, with over 20,600 citations out of a total of nearly 466,000. Following that were sources like NDR.de, a reputable German public broadcaster’s health content, and well-known medical reference sites such as MSD Manual.
While the study’s scope was limited to a specific region and time—December 2025—it raises universal concerns. If in a country with stringent health standards, AI Overviews still heavily depend on YouTube, it suggests that the problem could be widespread.
AI and health expert Hannah van Kolfschooten from the University of Basel commented that these findings reveal inherent risks within the design of Google’s AI summaries. She emphasized that the reliance on YouTube’s popularity rather than medical authority points to a systemic issue—one where visibility and engagement often overshadow factual accuracy.
Google’s representatives defend their approach, arguing that the most cited YouTube videos are predominantly from credible medical channels, with 96% of the top 25 videos originating from certified healthcare sources. Still, researchers caution that these examples constitute less than 1% of all cited videos, which means most YouTube content referenced could lack proper medical oversight.
So, the question remains: How much can we trust Google’s AI-driven health advice when the sources that heavily influence it include a platform famed for viral videos and user-generated content, not necessarily validated medical information? Is there a risk that influencer-led or unverified videos could soon dominate the health landscape online, even if they appear at the top of our search results?
What do you believe? Should there be stricter controls on what sources Google’s AI pulls from for health information? Or is this a broader glitch in the digital information age we must learn to navigate? Share your thoughts and join the conversation—because in health, misinformation isn't just inconvenient; it can be dangerous.
