01/12/2026
AI is increasingly showing up in healthcare, and this month both OpenAI and Anthropic announced new health-focused platforms aimed at supporting how people interact with health information and systems.
has introduced ChatGPT Health, a health-focused experience within ChatGPT that lets users choose to connect medical records and wellness app data for more personalized responses. Through a partnership with b.well Connected Health, access to health data is consent-based and secured with added protections. OpenAI states that health conversations are kept separate from other chats, are not used for model training, and are designed to support health understanding and care preparation, not diagnosis or treatment.
Anthropic has also entered the healthcare space with Claude for Healthcare, which is primarily aimed at health systems, payers, and life sciences organizations. The platform is built to work with protected health information and connects to commonly used medical and scientific resources. Claude for Healthcare is intended to support administrative, operational, and research workflows, such as documentation, billing, and prior authorization, while also allowing individuals to review and summarize their own health information when authorized.
Although both companies are entering the healthcare space, their strategies differ. OpenAI is keeping its consumer health features separate from tools built for hospitals and health systems, while Anthropic is combining consumer and healthcare organization features within a single platform. In both cases, the emphasis remains on user consent, privacy protections, and positioning AI as an assistive tool rather than a substitute for medical judgment.
From a dermatology perspective, it’s clear how quickly AI is becoming part of healthcare, but its limitations matter. I’ve seen patients come in after using AI tools to self-diagnose a rash, reassured it was something benign like granuloma annulare, only for a biopsy to later suggest a diagnosis such as cutaneous T-cell lymphoma. Examples like this highlight why these tools must remain supportive, not definitive, and why real clinical evaluation, examination, and pathology remain essential.