Google’s AI-generated responses to users’ health queries aren’t very helpful—yet

The trend: Consumers generally find that AI-generated responses to their online health queries are only somewhat reliable, according to a new survey from the Annenberg Public Policy Center (APPC) of the University of Pennsylvania. Over 1,600 US adults were polled in April.

Digging into the trend: 71% of respondents said they go to Google or Bing at least occasionally when they have a question about a medical symptom or condition. Among this sample, nearly two-thirds (65%) say they have seen AI-generated responses at the top of search engine results.

The quality of AI-generated responses to users’ health searches is mixed.

  • 45% said they sometimes get the answer they were looking for.
  • 26% said they often do.
  • Just 4% report always getting the answer they need.

Why it matters: A key concern of brands and publishers in the AI search era is that users end their search sessions quickly and won’t click through to links.

  • For instance, Google search clickthrough rates (CTRs) for healthcare content dropped 34.5% from May to December last year, according to our Industry KPI data provided by Databox.
  • The drop in CTRs aligns with the May 2024 launch of Google’s AI Overviews.
  • AI Overviews more often appear when queries require authoritative answers, such as medical definitions in healthcare, per a May 2025 BrightEdge report.

You've read 0 of 2 free articles this month.

Get more articles - create your free account today!