Back
Technology

Google AI Overviews Face Scrutiny Over Health Information Accuracy and Sourcing

Generated on: Last updated: 3 sources
View source

Google's AI Overviews Face Scrutiny Over Health Information Accuracy and Sourcing

Google's AI Overviews feature has recently come under intense scrutiny regarding the accuracy of its health-related information and its sourcing practices. Investigations have uncovered instances where the AI summaries provided advice contradicting medical consensus, while a separate study indicated a frequent reliance on YouTube for health queries. These findings raise significant questions about the platform's methodology and the potential impact on public health.

Identified Inaccuracies in Health Information

An investigation highlighted several cases where Google's AI Overviews presented inaccurate health information, drawing concern from medical experts and health organizations:

  • Pancreatic Cancer Advice: One summary incorrectly advised individuals with pancreatic cancer to avoid high-fat foods. Medical experts noted this recommendation contradicts standard advice and could negatively affect patient health and treatment eligibility.
  • Liver Function Tests: Information on liver function tests was described as inaccurate, potentially leading individuals with serious liver conditions to misinterpret their health status.
  • Women's Cancer Screening: Information regarding women's cancer tests, specifically for vaginal cancer, was found to be incorrect. A Pap test was incorrectly listed as a diagnostic tool. Experts stated this could deter individuals from seeking appropriate medical evaluation.
  • Mental Health Information: AI Overviews related to conditions like psychosis and eating disorders were described as potentially misleading or lacking crucial context.

Health organizations and professionals have expressed significant concerns. The Patient Information Forum noted the risk of inaccurate information appearing prominently, while Marie Curie highlighted potential harm to vulnerable individuals. Pancreatic Cancer UK, the British Liver Trust, Eve Appeal, and Mind each pointed to specific risks associated with the inaccuracies in their respective fields, including compromised treatment, delayed care, and avoidance of professional help.

Sourcing Practices Under the Microscope

A separate study conducted by SE Ranking analyzed over 50,000 German-language health queries performed in Berlin. This study indicated that Google's AI Overviews cite YouTube more frequently than dedicated medical websites for health-related responses.

YouTube's Prominence

  • YouTube was identified as the most cited source, accounting for 4.43% of all AI Overview citations in the study.
  • Researchers noted that no hospital network, government health portal, medical association, or academic institution achieved comparable citation numbers.
  • SE Ranking researchers expressed concern that YouTube, as a general-purpose video platform, allows content uploads from various sources, including those without medical training.

An AI, health, and law researcher from the University of Basel, not involved in the study, commented that the findings suggest the risks posed by AI Overviews for health are "structural." The researcher added that the heavy reliance on YouTube implies "visibility and popularity, rather than medical reliability, is the central driver for health knowledge."

Other Cited Sources

Other top cited domains included NDR.de (German public broadcaster), Msdmanuals.com (medical reference site), Netdoktor.de (a consumer health portal), and Praktischarzt.de (a career platform for doctors).

Google's Official Response

Google has addressed the concerns, responding to both the identified inaccuracies and the sourcing practices.

Regarding the health inaccuracies, Google stated that many of the examples provided were "incomplete screenshots." The company maintains that its AI Overviews often link to "well-known, reputable sources and recommend seeking out expert advice." Google also stated it significantly invests in the quality of AI Overviews, particularly for health topics, and that the majority provide accurate information, with an accuracy rate consistent with other search features like featured snippets. The company added that it takes action under its policies when AI Overviews misinterpret web content or miss context.

Concerning the study on sourcing, Google stated that AI Overviews are designed to surface high-quality content from reputable sources, regardless of format. The company noted that various credible health authorities and licensed medical professionals create content on YouTube. Google also suggested that the study's findings, based on German-language queries in Germany, might not be generalizable to other regions. While Google indicated that 96% of the 25 most cited YouTube videos in the study were from medical channels, researchers cautioned that these videos constitute less than 1% of all YouTube links cited by AI Overviews in health contexts.

Broader Context and Study Limitations

These findings contribute to broader discussions regarding the reliability of AI-generated data. Previous reports have raised concerns about inaccurate financial advice from AI chatbots and issues with AI summaries of news content.

The SE Ranking study acknowledged limitations, including that it was a one-time snapshot conducted in December using German-language queries. Researchers noted that results could vary over time, by region, and based on question phrasing.