Google shutters AI health feature
Google pulled an AI health search feature after criticism that it surfaced misleading, non-expert advice—an expensive reminder that unvetted AI can erode trust in health contexts. The takedown underscores that accuracy, expert validation, and transparent sourcing are now minimum viability requirements for health-facing AI. (breitbart.com)
Google confirmed it has discontinued the AI Search feature “What People Suggest,” a tool that surfaced crowdsourced health tips, according to reporting published March 16, 2026. (theguardian.com) “What People Suggest” was unveiled at Google’s health event in March 2025 and used AI to summarize lived-experience posts from forums including Reddit, Quora and X into topical guidance for mobile searches. (techcrunch.com) The takedown follows a January 2026 Guardian investigation that prompted Google to pull some AI Overviews for medical queries after experts flagged dangerous inaccuracies—examples cited included misleading dietary advice for pancreatic cancer patients. (techcrunch.com) Google told The Guardian the feature was “turned down months ago as part of a broader simplification” of Search and said the removal was not related to safety or quality concerns. (theguardian.com) At Google’s Check Up event on March 17, 2026, the company announced a $10 million investment to train clinicians in AI literacy and new Fitbit upgrades that include medical-record linking and a stated ~15% sleep-staging accuracy improvement for public-preview users. (blog.google) Multiple outlets reported Google quietly deprecated “What People Suggest” roughly a year after launch and did not issue a standalone public announcement about the feature’s removal. (ghacks.net)
