OpenAI has introduced a new Health feature in ChatGPT that enables users to upload medical records and connect fitness data from Apple Health, aiming to provide personalized health insights. However, this development has sparked concerns from experts about data privacy risks and the accuracy of AI-generated medical advice.
What Happened
On July 25, 2026, OpenAI announced the rollout of “Health in ChatGPT,” a feature allowing users to upload medical documents such as lab results and medication lists. It can also access health and fitness data through Apple Health with the user’s permission. The platform promises to compare new results with prior tests, summarize health changes since the last appointment, and explore correlations between activity and routine. OpenAI stated that 300 million people ask ChatGPT health-related questions each week, prompting this new dedicated space to better assist users.
Key Facts
OpenAI emphasized multiple safeguards for the feature, including additional encryption and confirmation prompts before performing connected actions like sending emails. The company assured that medical records uploaded will not be used to train AI models or targeted for advertising purposes. Despite these measures, the platform cautions users that ChatGPT is not a doctor and cannot replace professional medical care, diagnosis, or treatment.
The rollout follows recent incidents underscoring AI health risks: a Florida man filed a lawsuit claiming ChatGPT’s advice nearly caused his death from a pulmonary embolism, and a California teenager died after following AI-generated advice on consuming a dangerous substance. OpenAI responded to these incidents acknowledging AI’s limitations and removed problematic AI versions from public access.
What This Means
This new feature highlights a growing trend of individuals turning to AI for health guidance amid widespread dissatisfaction with traditional healthcare access. By aggregating personal medical data, ChatGPT aims to provide users with contextualized insights that might supplement healthcare providers’ assessments. However, the risks are considerable: users must balance convenience against potential privacy breaches and misinformation.
The reliance on AI for medical advice involves critical trade-offs. While AI might surface patterns missed by human clinicians, it can also produce inaccurate or harmful recommendations—a risk compounded for vulnerable populations such as the elderly or cognitively impaired who may not fully grasp these limitations. Moreover, the extensive sharing of sensitive health data with an AI tool raises longstanding privacy concerns, particularly over how data may be stored, protected, or used in the future despite vendor assurances.
Users should be cautious about submitting personal medical information to ChatGPT and should never substitute AI advice for professional care. This development serves as a reminder that AI health tools must be complemented with robust privacy policies and clear warnings to prevent misuse, misunderstanding, and harm.
Background
Previously, ChatGPT’s use for health-related queries has drawn attention due to instances of inaccurate or dangerous advice, as illustrated by the fatal case in California and the near-fatal lawsuit in Florida. These episodes have pressured OpenAI to refine its platform, adding disclaimers and disabling versions deemed too risky. The launch of a dedicated health feature seeks to formalize the interaction and introduce new safeguards, yet ongoing apprehension about AI in healthcare persists.
What Remains Unclear
It remains unknown how widely users will adopt the Health feature or how effectively OpenAI’s encryption and consent mechanisms will protect user data in practice. Questions about long-term data retention and policy changes remain unanswered, especially given past regulatory warnings regarding AI companies quietly modifying terms of service. OpenAI has not specified whether ChatGPT can reliably recognize emergencies or prompt users to seek immediate medical attention.
Sources
This article is based on reporting and publicly available information from the following sources:
Read more Artificial Intelligence stories on Goka World News.