The news: Atlantic Health system partnered with digital health company K Health to launch PatientGPT, a chatbot that integrates with the organization’s MyChart patient portal. The tool pulls from each patient's medical profile inside MyChart to personalize responses, direct patients to the right level of care, and loop in a clinician when needed. PatientGPT is launching in a beta phase and will roll out gradually to patients across New Jersey over the next few months.

Why it matters: Patients already use AI for health questions, but they’re primarily turning to general-purpose chatbots—providers want to be their trusted source of health information.

• AI for health will be the fastest-growing AI use case we measure in 2026.
• We expect US genAI health users, those who use genAI at least monthly for information, insights, or support related to their physical or mental health, to surge 37% YoY to 46.1 million, or 13.6% of the population, per EMARKETER forecasts.
• The caveat: Most people turn to mainstream AI tools like ChatGPT or Gemini for health queries (versus healthcare provider-developed AI tools). OpenAI reported earlier this year that 25% of its global weekly active users submit a healthcare-related prompt each week.

More healthcare provider organizations are building AI chatbots on top of patient records. K Health launched PatientGPT products with Hartford HealthCare and Novant Health this year, while Amazon added a Health AI assistant to the One Medical app that uses members’ records to explain lab results and flag when a doctor is needed.

By grounding AI in patients’ internal data and health histories, providers aim to deliver more private, credible, and personalized answers than public-facing chatbots. And they have trust on their side: Consumers are 3X as likely to trust an AI agent in a doctor’s secure patient portal as one on a public chatbot or general health website (60% versus 20%), per a June Salesforce report.

Implications for healthcare providers and AI companies: Health systems recognize that the digital patient experience is becoming a more important front door as consumers turn to AI chatbots for health guidance. Especially because in some cases, consumers use AI-generated health advice to skip or switch doctors. That raises the stakes for providers to keep patients within their own digital ecosystems and maintain greater influence over where they seek care, what information they trust, and when they engage with a clinician.

Still, patient-facing AI adoption remains limited as health systems weigh its potential against the risk of inaccurate or harmful advice. Making these tools widely used and valuable will likely require a more cautious rollout than consumer AI companies have pursued, including rigorous testing and validation of AI outputs for accuracy and safety, transparent marketing, and clear communication of the risks and benefits before patients use them.