Since early 2024, Google has been aggressively pushing its AI search overhaul, especially the “AI Overview” feature. This feature was originally hailed as the next evolution of the search experience, replacing the meaning explanations and summary displays that had been used for years. However, as user pushback intensified, Google recently quietly pulled one of its AI features related to health information, sparking considerable debate.
Google decided to cancel its AI medical advice feature
This terminated feature was an AI plugin called “User Recommendations” that Google was testing in search. Its concept was to automatically aggregate “collective wisdom” from internet forums, social platforms, and other public discussion boards when users searched for health symptoms, presenting the information as an easy-to-read AI-generated health advice summary. In other words, when you searched for “what to do for a sore throat” or “causes of dizziness,” Google would actively compile strangers’ personal experiences and present them to you in a seemingly professional format.
At first glance, this feature seems quite convenient. After all, when many people experience unexplained symptoms, their first step is often not to see a doctor, but to search online and browse experiences shared by others. Google’s idea is: since everyone is going to read these contents anyway, why not have AI organize them into a clearer version. However, this is precisely where the problems begin to emerge.
According to theThe GuardianAccording to reports, Google has quietly stopped testing this feature.Android Authority It also points out that while “compiling netizens’ experiences” may sound considerate, it raises a fundamental question: should search engines endorse strangers’ medical advice, or even package it as a seemingly reliable source of information?
The purpose of this feature is to help users quickly find experiences from others with similar symptoms, especially those with rare diseases or difficult-to-diagnose conditions. However, these contents are essentially “online comments” and not professional medical opinions. AI cannot assess commenters’ backgrounds, health status, age, or medical history, nor can it verify whether the information is accurate. Once these decontextualized suggestions are presented by AI in a formal tone, users may mistakenly assume they carry a certain level of credibility.
More importantly, Google’s recent AI performance in the medical information field has faced heavy external scrutiny. Previous investigations have found that Google’s AI health summaries sometimes provide incomplete, misleading, or even dangerous advice. For instance, some suggestions lack necessary background explanations, while others contradict official guidelines from medical institutions. Even when the content itself contains no errors, AI may overlook the fact that “health advice needs to be tailored to the individual”—which is precisely one of the most important principles in healthcare. Additionally, although Google includes disclaimers advising users to “consult a qualified healthcare professional” in its AI responses, this text is not always placed in a prominent position. For some users, the AI’s tone and presentation make it appear more authoritative than it actually is, which has further deepened external concerns.
In response to these criticisms, Google stated that removing the “User Suggestions” feature was not due to accuracy issues, but rather because the company is undertaking a larger-scale search experience simplification plan. In other words, Google believes this feature does not align with its future search direction, so it chose to terminate the testing.
Currently, AI in Google Search will no longer automatically compile crowd-sourced medical advice from across the web. Users who still want to see discussions on forums or social media can certainly search for them themselves, but Google will no longer organize this content using AI, nor will it package this content as “health advice summaries.” This change reflects a larger trend: the application of AI in medical information remains fraught with challenges. Although AI can quickly organize large amounts of data, medical information involves individual differences, professional judgment, and high stakes—far beyond what simple text compilation can resolve. Google’s rollback, in a way, also serves as a reminder to tech companies that in the health sector, speed and innovation can never supersede safety and accuracy.
Source: KOCPC Chinese