OpenAI has introduced a new feature called ChatGPT Health, which allows users to connect their medical and wellness records to the AI chatbot. The feature is designed to help individuals manage their health by providing personalized insights based on their personal data. However, concerns have emerged about the accuracy and reliability of the AI’s responses, as the system has been shown to generate incorrect or fabricated medical information—even when presented with real patient data. The tool enables users to upload health records such as lab results, medication lists, and doctor’s notes, which are then processed by the AI to answer questions about symptoms, treatment options, and health trends. OpenAI claims the feature is intended to support users in understanding their health data and preparing for conversations with healthcare providers. Despite these intentions, experts have raised red flags. Independent testing has revealed that ChatGPT Health sometimes invents medical details, such as fake diagnoses, non-existent test results, or incorrect drug interactions. In some cases, the AI has produced responses that contradict actual medical records or suggest treatments that are not supported by clinical evidence. OpenAI has acknowledged the risks and emphasized that the feature is not a substitute for professional medical advice. The company states that all responses are generated based on patterns in data and should be reviewed by a qualified healthcare provider before any decisions are made. It also notes that the system is still in a limited testing phase and is not yet available to the general public. The introduction of ChatGPT Health comes amid growing interest in using AI to improve health care access and personalization. However, the potential for AI to misinterpret or fabricate health information underscores the need for strong safeguards, transparency, and regulatory oversight. As AI becomes more integrated into health care, the balance between innovation and patient safety remains a critical challenge.