The advent of generative artificial intelligence (AI) has substantially altered — both positively and negatively — the ways in which people seek and obtain health-related information. People increasingly turn to large language models to make sense of their well-being, using them to interpret symptoms, plan diets, manage anxiety, and coach on sleep or fitness1,2. However, although the clinical AI community has rapidly coalesced around rigorous reporting guidelines for clinical trials3, there remains no equivalent, widely shared roadmap for consumer-facing applications. The recent launch of dedicated platforms like OpenAI’s ChatGPT Health and Ant Group’s AQ underscores this urgency.

It is crucial to distinguish the specific domain of these interactions. ‘Public health’ typically refers to population-level policy and epidemiology and ‘clinical medicine’ focuses on specialized diagnostics and treatment protocols, but the rise of large language models has created a third, distinct ecosystem. We term this domain ‘general health’: the vast, unmediated space of everyday health-information seeking. In this context, users independently query AI for lifestyle optimization, symptom interpretation and medical literacy. These tasks require safety and accuracy at scale, yet they differ fundamentally from the rigid binary outcomes of clinical diagnostics or the aggregate statistics of public health.

Enjoying our latest content?
Log in or create an account to continue

  • Access the most recent journalism from Nature's award-winning team
  • Explore the latest features & opinion covering groundbreaking research
Access through your institution

or

Sign in or create an account
Continue with Google
Continue with ORCiD