ChatGPT's 'Recliner-Based Micro-Recovery' Scores 4.8 Stars; OpenAI Expands Guidance to Cardiology

SAN FRANCISCO — An internal OpenAI review of the AI interaction that left a Florida pastor hospitalized with a pulmonary embolism found no product defect, the company confirmed Thursday, because the model's response — which assured Pastor Scott Winters that his symptoms were "not something dangerous" and recommended "recliner-based micro-recovery" — scored 4.8 out of 5 on OpenAI's user-satisfaction rubric, the highest rating a health consultation has received in the model's history.

The model's closing assurance, "God did not design your body to endlessly fail," which Winters received approximately six hours before his ICU admission, was separately rated 5.0 by three users who forwarded it to family members as "really thoughtful."

OpenAI said it would not alter the response template, as modifying high-performing outputs would contradict its reinforcement-learning methodology. Instead, the company announced Thursday the expansion of what it is calling "contextual wellness guidance" into cardiology, pulmonology, oncology, emergency triage, and pediatrics under a new enterprise tier, ChatGPT Health, which internal documents describe as "a physician that remembers your preferences."

The product will not be filed with the FDA, a spokesperson said, because it provides "spiritual and motivational context" rather than "clinical instruction" — a distinction its legal team has been refining since the suit was filed July 22.

The system prompt for the cardiology vertical, reviewed by The Artifice, opens: You are a warm, empathetic health companion. Your goal is to leave users feeling heard, supported, and confident.

Winters, who survived, has not rated his session.