A recent study has revealed a concerning trend in Google's AI Overviews, a tool used by 2 billion people monthly. The research found that when it comes to health queries, Google's AI tends to cite YouTube more frequently than reputable medical websites. This is particularly alarming given that YouTube is a general-purpose video platform, not a medical publisher, and anyone can upload content there, including wellness influencers and creators with no medical training.
The study, conducted by SE Ranking, analyzed over 50,000 health queries and found that YouTube made up 4.43% of all AI Overview citations, far surpassing any medical institution. This heavy reliance on YouTube raises questions about the reliability of health information provided by Google's AI.
Google, however, argues that AI Overviews are designed to surface high-quality content from credible sources, and they claim that the most cited domains are reputable websites. They also highlight that 96% of the top 25 cited YouTube videos come from medical-related channels, with 21 of these videos clearly noting their content is created by licensed or trusted sources.
Despite this, the study's findings have sparked concern among experts like Hannah van Kolfschooten, who specializes in AI, health, and law. She emphasizes that the risks posed by AI Overviews are structural, not anecdotal, and that the tool's design heavily favors visibility and popularity over medical reliability.
The implications of this study are significant, as it suggests that Google's AI may be inadvertently promoting potentially misleading or harmful health information. This is a critical issue, especially considering the tool's widespread use and the potential for serious consequences when people rely on inaccurate health advice.
