Google AI Overviews Put Health Searches and Patients at Risk

When a search result starts sounding like a diagnosis, the stakes change fast. 

Google’s AI Overviews are under scrutiny after experts warned that the tool is surfacing misleading health information that could put patients at risk.

According to an investigation by The Guardian, the generative AI summaries shown at the top of Google Search have produced inaccurate and potentially harmful medical guidance across health topics, raising fresh concerns about trusting AI-generated health answers when users are most vulnerable.

Helpful at a glance, harmful in detail

Google’s AI Overviews are positioned above traditional search results, presenting concise summaries that can read like medical guidance rather than general information. Health experts warn that when those summaries are flawed, their prominence can magnify the consequences.

One example cited by health groups showed Google advising people with pancreatic cancer to avoid high-fat foods. Anna Jewell, director of support, research, and influencing at Pancreatic Cancer UK, said the advice was incorrect and could leave patients too weak to tolerate chemotherapy or surgery.

“If someone followed what the search result told them, they might not take in enough calories, struggle to put on weight, and be unable to tolerate potentially life-saving treatment,” Jewell said.

Other summaries oversimplified diagnostic information. Searches about liver blood tests returned “normal” ranges without accounting for factors such as age or sex. Pamela Healy, chief executive of the British Liver Trust, warned this could falsely reassure people with serious liver disease and delay follow-up care.

Women’s health searches brought similar concerns. A query about vaginal cancer incorrectly listed a Pap test as a diagnostic tool. Athena Lamnisos, chief executive of The Eve Appeal, said the error could cause people to dismiss symptoms after a clear cervical screening result, delaying diagnosis of a rare but serious cancer.

How AI summaries can change patient behavior

Health charities say the deeper concern is how people respond after seeing these summaries. Clear, confident answers can shorten the distance between a search and a health decision, especially when users are anxious, unwell, or acting on behalf of someone else.

Stephanie Parker, director of digital at end-of-life charity Marie Curie, points to the emotional context behind many health searches. “People turn to the internet in moments of worry and crisis,” she said, warning that information that is inaccurate or stripped of nuance can “seriously harm their health.”

For organizations focused on patient information standards, trust is the central issue. Sophie Randall, director of the Patient Information Forum, said the placement of AI-generated summaries increases the chance they will be treated as definitive rather than provisional.

“The examples show Google’s AI Overviews can put inaccurate health information at the top of online searches,” Randall noted, adding that this placement increases the risk of people acting on what they see instead of consulting clinicians or attending follow-up care.

‘Very dangerous advice’

The Guardian found AI Overviews returning summaries for conditions such as psychosis and eating disorders that experts say lacked essential context.

Stephen Buckley, head of information at mental health charity Mind, said some of the AI-generated summaries offered “very dangerous advice” or were “incorrect, harmful or could lead people to avoid seeking help.” In other cases, the responses stripped away nuance that is critical when discussing serious mental health conditions.

Buckley also said some summaries directed users to inappropriate sources or presented information in ways that could reinforce biases or stigmatizing narratives. In mental health searches, he cautioned, even small omissions can influence whether someone seeks professional care or chooses not to seek help at all.

Google disputes the scale of the problem

According to Google, the vast majority of AI Overviews deliver accurate, helpful information. A Google spokesperson said many of the examples highlighted were based on incomplete screenshots and that, where the company could assess them, the summaries linked to “well-known, reputable sources” and encouraged users to seek expert advice.

The tech giant also said the accuracy of AI Overviews is comparable to long-standing search features such as featured snippets, and argued that it invests heavily in quality controls, particularly for health-related topics. When summaries misinterpret content or miss context, Google said it takes action under its existing policies.

Search has long been a gateway to medical information, but AI summaries are changing how far users dig for answers. Health experts warn that fewer clicks and faster conclusions may come at a cost when nuance matters most.

As tech companies rethink how people interact with AI, OpenAI is preparing a voice-first device that could move AI away from screens and into everyday listening.