Google removes AI summaries for certain medical queries

Following a Guardian investigation that found Google AI Overview provided misleading information for certain health-related queries, the company appears to have removed AI Overview for some of those queries.

For example, the Guardian initially reported that when users ask, “What is the normal range for a liver blood test?” they are presented with a number that does not take into account factors such as nationality, gender, ethnicity, or age, potentially leading them to believe they are healthy when their results are not.

The Guardian now states that the AI ​​summary has been removed from the results of ‘What is the normal range for a liver blood test?’ and ‘What is the normal range for a liver function test.’ However, we found that variations of queries such as “lft reference range” and “lft test reference range” can still lead to AI-generated summaries.

When I tried these queries this morning, a few hours after the article was published in the Guardian, none of them showed an AI summary, but Google still gave me the option to ask the same queries in AI mode. In some cases, Guardian articles about deletions were actually the top results.

A Google spokesperson told the Guardian that the company “does not comment on individual deletions within search” but said it was working on “broad improvements.” The spokesperson also said an in-house team of clinicians reviewed the questions raised by the Guardian and found that “in many cases the information is not inaccurate and is supported by a high-quality website”.

TechCrunch has reached out to Google for additional comment. Last year, the company announced new features aimed at improving Google Search for medical use cases, including improved summaries and health-focused AI models.

Vanessa Hebditch, director of communications and policy at the British Liver Trust, told the Guardian that the removal was “great news”, but added: “Our bigger concern about this is that Google is cherry-picking a single search result, which allows them to block AI overviews, but they’re not addressing the larger issue of AI overviews for health.”