[CLSTR]
< Back to situation

[REVISION HISTORY]

OpenAI lawsuit and ChatGPT Health rollout under scrutiny

Updated 3 times since CLSTR started tracking revisions of this situation.

What changed

2026-08-04 11:25 UTC → 2026-08-23 06:04 UTC · added removed

On July 21 21, 2026, Florida pastor Scott Winters filed a civil suit alleging that ChatGPT‑4o’s reassurance that his dizziness against OpenAI and groin pain were not serious CEO Sam Altman in San Francisco Superior Court. The lawsuit alleges that a July 2025 interaction with ChatGPT-4o led him to delay care and suffer postpone medical care, resulting in a massive pulmonary embolism. OpenAI announced The complaint claims the same day that its new “Health in ChatGPT” feature would let U.S. adults 18+ link Apple Health, electronic‑medical‑record systems AI suggested specific medications and wearable data to the chatbot. The company emphasized physical postures that user data are encrypted, under user control, and not used caused him to train the model, and reiterated collapse, alleging that the service OpenAI prioritized profits over safety. The plaintiff is not seeking damages and a substitute for professional mechanism to automatically terminate conversations when urgent medical care. The rollout began assistance is required. Simultaneously, OpenAI launched its “Health in ChatGPT” feature on July 23 2026 and was made available 23, 2026. Available to U.S. adults 18+ across free, Plus Plus, and Pro plans, running on the GPT‑5.6 service utilizes the GPT-5.6 Luna model for paid tiers (GPT‑5.5 Instant for free users). OpenAI’s tiers. It allows users to link Apple Health, electronic medical records, and data from systems like One Medical and Function Health. While Vice President of Health Product Ashley Alexander claimed the models “reason at a level better than a clinician,” a statement later Health Lead Karan Singhal tempered this by the health lead, and emphasizing the service is not for diagnosis or treatment. OpenAI’s terms of service explicitly prohibit diagnostic or treatment such use. Weekly health queries now exceed 300 million worldwide. U.S. healthcare professionals warned that have raised concerns regarding privacy, consent, and the potential for patients might to treat AI‑generated AI responses as medical advice, raising privacy, consent, data‑security and patient‑provider‑relationship concerns. advice. Scientific studies offered conflicting assessments of AI performance in clinical settings: remain divided: a Nature Medicine paper found general‑purpose suggested general-purpose models (ChatGPT, Claude, Gemini) outperformed a specialized model (OpenEvidence), ones, while a later JAMA study rated OpenEvidence higher on accuracy, utility and verifiability. The lawsuit also references a July 2025 interaction with ChatGPT‑4o, underscoring found the tension between promotional claims specialized model OpenEvidence to be superior in accuracy and legal restrictions. utility. Calls for transparent, independent independent, transparent evaluation of AI health these tools continue persist as the rollout proceeds. continues.

Versions

  1. 2026-08-23 06:04 UTC OpenAI lawsuit and ChatGPT Health rollout under scrutiny
  2. 2026-08-04 11:25 UTC OpenAI lawsuit and ChatGPT Health rollout under scrutiny
  3. 2026-07-31 15:50 UTC OpenAI lawsuit and ChatGPT Health rollout under scrutiny
  4. 2026-07-26 15:18 UTC OpenAI lawsuit and ChatGPT Health rollout

Only revisions since CLSTR began indexing content versions appear here. Select a version to see what changed compared to the one before it.