Recently, OpenAI officially launched a new feature called “ChatGPT Health” in its AI product ChatGPT, marking a significant step for the company into the healthcare sector. This feature aims to provide users with analysis of physical examination results, guidance for medical appointments, and personalized diet and exercise plan recommendations. OpenAI has clearly stated that the service is designed to offer helpful health information support and does not provide formal medical diagnoses. To achieve more comprehensive health insights, the feature can connect to users’ electronic health records, wearable devices (such as Apple Health), and various health management applications. Currently, users can register for access through a waitlist, and broader availability is expected in the coming weeks.
With advancements in artificial intelligence, the ability to parse complex health data to provide personalized advice has attracted many technology companies to the lucrative healthcare market. However, the handling of highly sensitive personal health data by such services has also raised widespread concerns about privacy protection and data security. OpenAI revealed that over 200 million users consult ChatGPT on health-related issues every week. To enhance the professionalism and reliability of its services, the company has consulted with more than 260 doctors over the past two years to optimize its health features. Additionally, it plans to implement enhanced privacy protections for health-related conversations, isolating them from other usage scenarios.
The company emphasized that its AI health assistant is designed to assist rather than replace the judgment of medical professionals. Fei-Fei Li, CEO of OpenAI’s Application Division, noted that the strength of AI lies in its lack of time and energy constraints. It can deeply understand users’ situations and explain them in an easily comprehensible manner, thereby addressing some limitations of traditional medical consultations.
While expanding into healthcare applications, OpenAI is also focusing heavily on enhancing its audio AI capabilities to lay the groundwork for future personal AI devices centered around voice interaction. It is reported that the company is developing a consumer hardware device that primarily relies on auditory rather than visual screen interaction.
Currently, ChatGPT’s voice features and text responses are based on different models. The company’s internal evaluation found that existing audio models fall short in terms of accuracy and response speed. To address this, OpenAI has pooled resources over the past two months to tackle key technical challenges in audio modeling. Progress has been made in technical improvements, with the new model architecture capable of generating more natural and emotionally expressive speech. It also features more human-like interaction abilities, such as concurrent speech with users and handling interruptions. The audio model is slated for official release in the first quarter of 2026.
OpenAI’s hardware vision is not limited to a single product but encompasses a product line that includes smart glasses and screenless smart speakers. The core concept is to create an “always-on assistant” capable of proactively sensing the environment and user needs, providing continuous assistance through audio and visual means with user authorization. To support this long-term strategy, OpenAI completed the acquisition of the renowned design firm io in early 2025. It is also advancing multiple fronts simultaneously, including supply chain development, industrial design, and core model research and development, in preparation for its first hardware product, which is expected to launch at least a year from now.
