OpenAI Reports 600,000 Weekly Health Queries from Hospital Deserts, 70% After Hours

OpenAI’s health‑focused usage data, disclosed by Chengpeng Mou, the company’s head of business finance, shows that roughly 600,000 of the two million weekly health‑related messages sent by U.S. users originate from “hospital deserts”—areas where the nearest emergency facility lies at least a half‑hour away. The breakdown, which comes from anonymized usage logs, also reveals that 70 percent of those queries are submitted outside conventional office hours, underscoring a demand for round‑the‑clock medical information in regions where physical access is limited (The‑Decoder, Apr 6 2026).

Mou’s remarks followed a viral X post by Simon Smith, who described how his family leveraged a shared ChatGPT workspace to collate advice from multiple clinicians while caring for his ailing father. Smith’s anecdote, which sparked a broader conversation about the platform’s role in patient decision‑making, prompted OpenAI to highlight that such collaborative use cases are “not edge cases,” according to Mou. The implication is that a growing segment of the public is treating the chatbot as a de‑facto health liaison, especially when traditional channels are unavailable or inconvenient.

The figures arrive at a moment when OpenAI is intensifying its push into the U.S. healthcare ecosystem. In recent weeks the firm rolled out a dedicated health section inside ChatGPT and has been courting hospital systems to embed the model in clinical workflows. By quantifying the volume of after‑hours queries from underserved locales, OpenAI is building a data‑driven case for partnerships that could position its technology as a triage adjunct or a source of preliminary guidance—functions that could alleviate pressure on overburdened emergency departments and urgent‑care clinics (The‑Decoder).

Analysts note that the 600,000‑query metric, while sizable, represents only a fraction of the overall health‑insurance‑related traffic OpenAI processes—about 30 percent of the two million weekly health messages. Nonetheless, the concentration of after‑hours activity suggests a structural gap in the U.S. health‑service delivery model, where many patients lack immediate access to qualified professionals. If OpenAI can demonstrate clinical safety and regulatory compliance, its data could become a lever for hospitals seeking to extend virtual care into the night shift without incurring the full cost of staffing additional personnel.

OpenAI’s disclosure also raises questions about data privacy and the limits of AI‑driven advice. The company stresses that the usage figures are anonymized, but the broader industry debate about the appropriate role of large language models in medical decision‑making remains unresolved. As OpenAI continues to embed its chatbot in hospital settings, regulators and providers will likely scrutinize how the platform handles sensitive health information and whether it can reliably differentiate between informational queries and situations that demand professional intervention. The 600,000 weekly queries from hospital deserts thus serve as both a market signal and a cautionary benchmark for the next phase of AI integration in American health care.