General-purpose AI assistant (health use)
ChatGPT / ChatGPT Health
ChatGPT is the largest patient-directed general AI channel for health questions, with OpenAI reporting more than 300 million people asking health questions weekly. Health in ChatGPT reached general availability for U.S. adults on July 23, 2026 across Free, Go, Plus, and Pro plans on web, iOS, and Android, connecting records (with Epic and Oracle Health named in coverage), One Medical, Function Health, MyFitnessPal, and Apple Health. The refreshed draft flags the structural change since January: connected health information now informs all conversations rather than a sealed Health space, so the privacy promise rests on no-training and no-ad-targeting commitments rather than visible architecture. The service remains outside HIPAA, OpenAI defines its own safety benchmarks, and two lawsuits filed in May and July 2026 allege harmful guidance and seek a feature pause; those claims are pending allegations, not findings.
Seed classification is visible, but an evidence-backed public-source review is still pending.
Summary judgment · 75% toward patient-directed
Potentially agency-expanding, with platform-custody caveat
Patients choose and steer ChatGPT for their own health questions, now generally available with record grounding, no-training and no-ad-targeting commitments, and deletion controls. The custody caveat sharpened in July 2026 when OpenAI dissolved the isolated Health space so connected health information informs all conversations, and two pending lawsuits are testing recourse for allegedly harmful guidance.
Patient agency
How this tool changes agency
Explaining lab results, preparing appointment questions, interpreting wearable data, summarizing care instructions, and comparing insurance options are core advertised uses, developed with hundreds of physicians on evaluation rubrics; OpenAI concedes the model can still make mistakes.
General availability on July 23, 2026 removed the waitlist for U.S. adults on all plans, and connections are opt-in and disconnectable. Record connections remain U.S.-only, the EEA, Switzerland, and the UK are excluded, Android lacks a wearable-data path without Google Health support, and press reports a newer model (GPT-5.6 "Sol") for paid subscribers.
Patient-facing signals
Who does this AI serve?
Users choose the tool for their own questions and OpenAI markets Health as supporting, not replacing, care, but press coverage frames the launch as a push to become a hub for personal health data, with commercial app partnerships (Function, MyFitnessPal, Weight Watchers, Instacart, Peloton) built into the experience.
Can patients tell AI is involved?
ChatGPT is openly an AI product, health connections are explicit opt-ins, and OpenAI states the service is not intended for diagnosis or treatment; since July 2026 health context can inform any conversation rather than only a labeled Health space.
Can patients meaningfully choose?
General availability on July 23, 2026 removed the waitlist for U.S. adults on all plans, and connections are opt-in and disconnectable. Record connections remain U.S.-only, the EEA, Switzerland, and the UK are excluded, Android lacks a wearable-data path without Google Health support, and press reports a newer model (GPT-5.6 "Sol") for paid subscribers.
Can patients correct or challenge what the AI produces?
Users can delete chats and health memories, add custom instructions, and disconnect data sources, but there is no disclosed mechanism to correct model errors, audit an answer, or appeal harmful guidance; record errors must be fixed at the source provider, and two pending lawsuits are testing external recourse.
Does it help patients understand or act?
Explaining lab results, preparing appointment questions, interpreting wearable data, summarizing care instructions, and comparing insurance options are core advertised uses, developed with hundreds of physicians on evaluation rubrics; OpenAI concedes the model can still make mistakes.
Text findings
Who is left out or burdened?
Access widened; structural gaps moved rather than closed
General availability removed the waitlist, but record connections are U.S.-only, the EEA, Switzerland, and the UK remain excluded, the feature is 18+, Android users lack a Google Health path for wearable data, and paid subscribers reportedly get a newer model. Language, disability, and low-literacy support for the health experience are not publicly detailed.
What happens to patient data?
Strong stated commitments; the architectural boundary is gone
OpenAI commits that connected records, Apple Health data, and conversations using them are not used to train foundation models or target ads, with deletion controls. Since July 2026 health context flows across all conversations rather than an isolated space, the consumer service is not a HIPAA covered entity, retention specifics and legal-process exposure are governed by consumer terms, and no independent audit of the commitments was found.
Are the clinical boundaries clear?
Clear in wording, contested in litigation
OpenAI consistently states the service is not intended for diagnosis or treatment and urges professional consultation. Lawsuits filed in May and July 2026 allege harmful guidance in a fatal drug-interaction case and a pulmonary-embolism case and seek a feature pause; the claims are pending allegations, not findings.
Who defined what good looks like?
Vendor-defined, at unprecedented scale
OpenAI selected the physicians, designed the scenarios and rubrics, publishes its own benchmarks, and changed the data architecture unilaterally in July 2026. No independent, patient-partnered evaluation of the health experience was found.
Review provenance
Criteria
Same HugoScore CAIHL-derived criteria for every tool; public criteria are displayed from site/data/criteria.json, with fuller method notes in SCORING_FRAMEWORK.md.
Reviewer
AI-assisted public-source draft prepared in Claude (Cowork); no named human reviewer recorded.
AI / model
Claude Fable 5 (claude-fable-5).
Human review
No comprehensive human review has been completed or claimed for this draft profile.
Review date
2026-08-09
Limitations
No hands-on walkthrough, connect/disconnect/deletion flow testing, vendor interview, live privacy-flow deep read, independent model evaluation, or legal analysis of pending litigation.
Review method
Public-source review of OpenAI's January 7, 2026 launch announcement and July 23, 2026 general-availability announcement, plus credible press coverage of the rollout, usage scale, platform availability, and pending litigation, and HIPAA-status analyses. No hands-on product walkthrough, connect/disconnect/deletion flow test, vendor interview, privacy-policy deep read of the live Health flows, independent model evaluation, or legal analysis of pending lawsuits.
Research pending · Medium draft, official sources and credible reporting