After an investigation by The Guardian uncovered that Google’s AI-generated overviews were offering potentially misleading health information, the company has removed AI overviews for some health-related searches.
The investigation revealed that certain queries, such as “What is the normal range for liver blood tests?” returned AI-generated information that ignored critical factors like nationality, sex, ethnicity, and age. This omission could have caused users to mistakenly believe their results were healthy when they might not have been.
Following these findings, Google has removed AI overviews for the queries “What is the normal range for liver blood tests?” and “What is the normal range for liver function tests?” However, variations of these queries, such as “LFT reference range” or “LFT test reference range,” continued to show AI-generated summaries, at least initially.
Google’s Response to the Removal
After The Guardian‘s report, attempts to search those queries, several hours after the story was published, did not return AI overviews. Instead, users were given an option to enter “AI Mode” for the query. In some cases, the top result was an article from The Guardian discussing the removal itself.
A Google spokesperson told The Guardian that the company does not comment on individual search result removals but works to make broad improvements across its platform. The spokesperson also mentioned that an internal team of clinicians reviewed the disputed queries and found that while some of the information might not have been inaccurate, it was not comprehensive enough to reflect the nuances required for medical queries.
Ongoing Concerns About AI Overviews in Healthcare
While Google has made these changes, the issue raised concerns within the healthcare community. Vanessa Hebditch, the director of communications and policy at the British Liver Trust, responded to the removal by saying, “It’s excellent news,” but added that the broader issue of AI overviews in healthcare had not been addressed. She expressed concern that Google was simply removing overviews for specific queries rather than addressing the larger problem of AI-generated health information.
Last year, Google introduced new features intended to improve health-related search results, including the development of AI models and overviews designed for healthcare applications. However, Hebditch’s comments highlight the need for further scrutiny and reform of AI usage in medical searches.
As AI continues to play a role in healthcare information, the controversy surrounding Google’s AI overviews underscores the need for ongoing regulation and accuracy in medical searches. While Google has responded to the specific concerns raised by The Guardian, broader issues remain in how the company handles health-related queries on its platform.