Google takes down some AI summaries over health risk concerns

Google has removed some of its AI-generated health summaries following an investigation revealing that people were served inaccurate health information.

An earlier investigation by the Guardian revealed that some of the AI summaries on Google – snapshots of essential information generated by AI Overviews – provide misleading content that can put people’s health at risk.

In one such example, bogus information appeared when typing “what is the normal range for liver blood tests”. According to the Guardian, the returned overview did not account for any crucial personal details necessary when answering, such as nationality, sex, ethnicity, or age of patients. Such generalities could falsely suggest to ill patients that they are healthy, experts claim.

The Guardian says that Google has since removed AI Overviews for the search terms “what is the normal range for liver blood tests” and “what is the normal range for liver function tests”.

“We do not comment on individual removals within Search. In cases where AI Overviews miss some context, we work to make broad improvements, and we also take action under our policies where appropriate,” a Google spokesperson commented.

Sue Farrington, the chair of the Patient Information Forum, told the Guardian that while this is a good first step, much work is left to be done as there are “too many examples out there of Google AI Overviews giving people inaccurate health information.”

In another example shown in the investigation, Google wrongly advised those who suffer from pancreatic cancer to avoid high-fat foods. According to the experts cited by the Guardian, this is the opposite of the advice that should be offered in such a case.

The original investigation also mentioned searches related to women’s cancer tests, which offered “completely wrong information”. The publication says that AI Overviews continue to appear for some of these other examples related to cancer and mental health.

A spokesperson for Google at the time said that many of the shared health examples were “incomplete screenshots”, but based on the assessment, they linked “to well-known, reputable sources and recommend seeking out expert advice.”

The Guardian asked why these AI Overviews had not also been removed, to which a spokesperson responded:

“Our internal team of clinicians reviewed what’s been shared with us and found that in many instances, the information was not inaccurate and was also supported by high quality websites.”

Cybernews previously reported on misleading AI Overviews summaries returning for queries about the so-called Department of Government Efficiency (DOGE). In one example, an AI-provided summary described DOGE, which was briefly led by Elon Musk, as a “fictional entity” and a “conspiracy theory.”