Can You Trust AI for Health Questions?

AI-generated health answers now appear first in many online searches, but experts warn users to approach them cautiously. A 2024 KFF survey revealed 17% of U.S. adults—including 10% of those aged 65 and older—regularly ask AI chatbots like ChatGPT for health advice. While AI can provide useful summaries, it often "hallucinates" false or misleading medical information, making it difficult for users to detect inaccuracies. A 2025 study published in *NEJM AI* found people frequently rated AI-generated health responses as better than those from doctors, even when the AI answers were less accurate. Kisha Davis, MD, president-elect of the American Academy of Family Physicians, noted AI lacks personalized context, such as individual health history or family medical background. This can lead to advice based on broad research that may not apply to specific cases, particularly for underrepresented groups like women or minority populations. Privacy is another concern, as AI systems are not bound by the same strict privacy laws as medical professionals. User-provided health data could be exposed in breaches or repurposed by the companies operating these platforms. A 2026 study in *Nature* highlighted that while AI models could accurately diagnose medical conditions when given full case details, they performed poorly when users provided incomplete information during casual interactions. For reliable health information, experts recommend consulting official government sites like NHS Inform or MedlinePlus, reputable medical institutions such as the Mayo Clinic or Cleveland Clinic, or professional physician organizations like the American Academy of Family Physicians. These sources adhere to rigorous standards and provide evidence-based guidance tailored to individual needs. Ateev Mehrotra, MD, from Brown University School of Public Health, emphasized the need for caution, stating that while AI can assist, its limitations make it untrustworthy as a sole source for health decisions. Users should cross-reference AI answers with verified sources and, when in doubt, seek professional medical advice.