Google has reportedly removed certain AI-generated summaries for health-related searches after an investigation found that some of the information provided could be misleading.
The summaries, known as AI Overviews, appear at the top of search results and are designed to provide concise answers to user questions. A report by the Guardian newspaper found that several AI Overviews contained inaccurate health information, raising concerns about potential harm to users.
The investigation highlighted cases where the AI supplied numbers with little context in response to queries such as “what is the normal range for liver blood tests?” and “what is the normal range for liver function tests?” The results did not account for differences based on age, sex, ethnicity, or nationality. In some cases, Google’s AI extracted data from Max Healthcare, an Indian hospital chain in New Delhi, rather than providing verified global medical guidance.
Featured snippets, which also appear at the top of Google search results, differ from AI Overviews because they extract existing text from relevant websites rather than generating new content. However, the Guardian noted that even variations of liver test queries, such as “[liver function test] lft reference range,” continued to produce AI-generated summaries. Liver function tests measure proteins and enzymes in the blood to evaluate how well the liver is performing.
In one example, Google’s AI reportedly advised pancreatic cancer patients to avoid high-fat foods. Experts told the Guardian that such guidance could be dangerous, potentially increasing the risk of mortality among patients.
The Guardian’s findings come amid broader concerns about AI chatbots “hallucinating,” a term used to describe when AI systems generate false or misleading information due to incomplete or inaccurate data. Experts have warned that reliance on AI for medical information could pose serious risks if users interpret these responses as authoritative.
Euronews Next contacted Google to confirm whether AI Overviews had been removed from certain health queries but did not receive an immediate response. Google announced over the weekend that it would expand AI Overviews to Gmail, allowing users to ask questions about their emails and receive automated answers without searching through messages manually.
The development underscores ongoing tensions between AI innovation and accuracy, particularly in sensitive areas such as healthcare. As AI tools become more integrated into search engines and email platforms, experts emphasize the importance of verifying information with trusted medical sources and cautioning users against relying solely on machine-generated summaries.
