Artificial intelligence is no longer just lurking in the background of Bay Area tech offices. A new national health poll shows Americans are now asking chatbots what to do about their bodies, and in some cases, those answers are changing whether they see a clinician at all. The West Health–Gallup survey points to millions of adults who say AI advice led them to skip care, and Stanford Health Care’s chief data scientist recently went on local TV to break down what that really means for everyday patients.
Poll finds one in four used an AI tool; 14 million skipped care
The West Health–Gallup Center surveyed 5,660 U.S. adults between Oct. 27 and Dec. 22, 2025, and found that roughly 25% of adults had used an AI tool or chatbot for health information in the past 30 days, according to Gallup. Gallup reports that 14% of recent AI users said the advice led them to skip a provider visit in the prior month, a rate the poll projects will affect about 14 million adults. At the same time, only a small share of users, about 4%, say they “strongly” trust the accuracy of AI health information.
Stanford scientist breaks down risks and benefits
On Bay Area television, Dr. Nigam Shah, chief data scientist at Stanford Health Care, joined KTVU's “The Nine” to react to the poll and offer some local context. KTVU and Shah described AI as a potentially useful tool for getting ready for appointments or interpreting test results, while also warning viewers not to lean too hard on systems that can simply get things wrong at times. Shah’s role and ongoing work at Stanford help explain why health systems and clinicians are watching how these tools are adopted both locally and across the country; his profile is available at Stanford Medicine, and the station’s segment on the poll can be seen on KTVU.
What people are asking AI, and which tools they use
Gallup reports that recent users most often turn to AI for everyday topics such as nutrition and exercise, about 59% of recent users, physical symptoms, 58%, and medication side effects, 46%. General conversational systems like ChatGPT or Microsoft Copilot are the most commonly used tools. Many respondents said AI helped them feel more confident when talking with a clinician, yet Gallup also found a noticeable minority who believed AI had recommended unsafe guidance.
Access gaps, privacy and expert concerns
Reporting by the Associated Press highlights why some people turn to chatbots in the first place: cost, limited access to care and the convenience of getting quick answers outside business hours, especially among younger and lower-income respondents. The AMA’s president told the AP News that AI can sharpen patients’ questions, but “it is an assistant but not an expert,” and experts caution that privacy, accuracy, and the risk of misleading recommendations remain real concerns for patients and health systems.
Local takeaways for readers
For Bay Area patients and providers, the immediate takeaway is practical rather than futuristic: use AI to get organized before visits, double-check any sources cited by chatbots and treat machine-generated advice as background information instead of definitive clinical guidance. If an AI tool suggests skipping care for worrisome symptoms, experts say that is a signal to verify the advice with a clinician rather than assuming the chatbot’s answer is sufficient.
Where systems need to go next
Gallup’s data point to a broader policy and operational challenge. Health systems now face pressure to develop clear guidance for safe AI use, build patient-facing literacy around limitations and privacy, and identify where vetted AI can reliably augment care. As usage grows, those local steps will shape whether AI settles in as a helpful supplement or slides into an unsafe substitute for in-person care.
