In less than a year, Google launched and quietly buried a feature that used AI to serve up health advice from internet strangers. The tool, titled “What People Suggest,” aggregated community health perspectives from across the web and organized them using artificial intelligence. Three people with knowledge of the decision confirmed its removal, which Google later acknowledged while offering a muted explanation.
The feature was announced with fanfare at Google’s “The Check Up” health event in New York, where it was positioned as a forward-thinking step in personalized health information. Former chief health officer Karen DeSalvo penned a blog post explaining how the feature would help patients connect with lived experiences from people managing similar health conditions. Arthritis sufferers, for example, could discover how others with the same diagnosis approached exercise.
Despite the optimistic launch, the feature did not survive. Google’s spokesperson maintained that safety had nothing to do with its removal, attributing it to a routine streamlining of search page elements. But the explanation fell flat when Google failed to produce a clear, contemporaneous public notice about the change, instead pointing to a blog post that did not reference the feature.
The story unfolds against a backdrop of deepening scrutiny over Google’s AI health tools. An investigation published earlier this year found that Google’s AI Overviews were delivering false health information to billions of users. That finding prompted Google to disable AI Overviews for some medical searches, though critics noted the response was slow and incomplete.
As Google schedules its next health event to tout AI-powered medical innovations, the abrupt end of “What People Suggest” serves as a cautionary tale. Building credibility in the health AI space requires more than impressive launch events and bold promises. It demands consistent quality, transparent communication, and a willingness to publicly own and explain product failures.