By Pete Mercer
OpenAI’s latest platform aims to centralize medical records, wellness data and AI guidance in one place – raising both optimism about patient empowerment and concerns over accuracy, privacy and trust.
In January, OpenAI announced the launch of ChatGPT Health, which is an extension of the AI chatbot that will act as a hub where people can connect their wellness apps and medical records, allowing ChatGPT to help individuals better understand and interpret any recent test results, prepare for upcoming doctor appointments, navigate differentiating insurance options, and advise on diet and workout plans.
The idea is that your medical records and data from the health apps would work to inform ChatGPT Health’s answers to any questions, centralizing this information to make it accessible to the user. Any questions you ask ChatGPT Health will be “grounded in your own health information,” according to a statement from the company.
According to the release, ChatGPT worked with 260 licensed physicians across 60 countries of practice to evaluate model performance using real clinical scenarios to test and train the platform. This physician-led training built a foundation for ChatGPT Health, testing more than 600,000 model outputs spanning 30 areas of focus. With this input from providers and other benchmarks in place to make the platform more reliable, ChatGPT and OpenAI are working to “ensure that AI benefits all of humanity.”
This is a significant development, and it could indicate a huge shift in the delivery of healthcare. Depending on how you feel about the emergence of artificial intelligence and the intersection of data and healthcare, this is either exciting or a big red flag.
Using AI for health consultations
The idea of using ChatGPT for health questions is certainly not a new one. Axios reported that “more than 40 million people globally turn to ChatGPT daily for health information,” viewing the platform “as an ally in navigating their health care, according to analysis of anonymized interactions with ChatGPT and a survey of ChatGPT users by the AI-powered tool Knit.”
Patients are already using ChatGPT to ask questions about healthcare, whether it’s to get information about symptoms, ask questions about navigating health insurance, or seek “deeper” context about any of their healthcare issues. It makes sense that OpenAI and ChatGPT would take advantage of a huge opportunity to provide an extra service for their users, but it also comes with its fair share of challenges.
The use of AI in these scenarios is still largely divisive, which is in no small part due to the fact that ChatGPT can give the wrong information out to its users. A study published by the BBC and EBU in October 2025 found that “around 45% of AI news queries to ChatGPT, MS Copilot, Gemini, and Perplexity produce errors.” Worse than that, the information could even turn out to be deadly.
A report from The Transparency Coalition, an organization devoted to ensuring that AI technologies are developed to prioritize safety and transparency, found that seven new lawsuits were filed against OpenAI in November 2025. These suits included various allegations like wrongful death and assisted suicide after several notable cases of suicides of individuals who were heavily engaging with the platform. The argument is that these mental health crises were triggered by the contents of the conversations people were having with the chatbot.
How does this factor into ChatGPT Health? While this new feature to the platform might add a certain level of convenience and even a better understanding of complex patient information, it shouldn’t be treated as a provider. In fact, OpenAI and ChatGPT both argue that this new feature is not designed or intended to replace providers.
ChatGPT said the following in the announcement for ChatGPT Health: “[ChatGPT] Health is designed to support, not replace, medical care. It is not intended for diagnosis or treatment. Instead, it helps you navigate everyday questions and understand patterns over time – not just moments of illness – so you can feel more informed and prepared for important medical conversations.”
Security and privacy concerns
One of the biggest obstacles to launching a product that uses an AI platform is any anxieties over the security and privacy of the user’s sensitive health records. This new addition to ChatGPT exists in its own space, where any conversations, connected apps, or files that you have previously used are kept separate from the Health component of the platform.
While the healthcare conversations you have in the platform will stay contained within the context and confines of the Health space, the bot can use context from your non-Health chats, like any recent lifestyle changes to make the conversation more relevant. ChatGPT said in the media release, “Due to the sensitive nature of health data, Health builds on this foundation with additional, layered protections – including purpose-built encryption and isolation – to keep health conversations protected and compartmentalized. Conversations in Health are not used to train our foundation models.”
Time Magazine spoke to Dr. Danielle Bitterman, a radiation oncologist and clinical lead for data science and AI at Mass General Brigham Digital, about the launch of ChatGPT Health. While she wasn’t surprised by the announcement, she did urge users to exhibit caution when using the platform: “The most conservative approach is to assume that any information you upload into these tools, or any information that may be in application you otherwise link to the tools, will no longer be private.”
Dr. David Liebovitz, co-director of the Institute for Artificial Intelligence in Medicine’s Center for Medical Education in Data Science and Digital Health at Northwestern University Feinberg School of Medicine, has a positive outlook on the utility of the platform, but he’s also aware of the challenges with security and confidentiality.
He said, “Privacy and security and confidentiality guarantees are not covered when a consumer, no longer a patient with protected health information, but a consumer uploads information to a third party. So those issues need to be carefully reviewed and addressed. There are ways in the meantime to give snippets for assessment as opposed to comprehensive records. It remains a concerning aspect in terms of secondary use of this data that’s outside the control of the consumer who uploaded this information.”
Is this a good thing?
The big question here is whether any of this is a good idea. It’s a fair question – even with all of the security and privacy measures, it will be hard for some people to trust the platform with their sensitive, personal information. Despite the risks, Dr. Liebovitz thinks that ChatGPT Health is going to be a beneficial tool for patient safety.
“I feel that anything that further empowers patients will actually be helpful given a focus on patient safety throughout my career,” he said. “That is, patients will have access to expertise that continues to improve and (should in the future) well exceed a random internet-based search engine. This parallels the development and application of rules and now machine learning algorithms leading to generative AI supporting the workflow of physicians with tools and efficiency gains by providing expertise here for patients to assist with understanding their medical care.”
The reality is that platforms like this are a huge opportunity for patients to develop a better understanding of test results or a diagnosis, especially in those moments when the physician is unavailable to communicate. For Dr. Liebovitz, this is an inevitable step towards the democratization of healthcare. “We’ll wonder about the time we used to practice when we didn’t have additional safeguards throughout all steps of care.”
As for the scenarios where people are given inaccurate information from AI platforms, Dr. Liebovitz pointed out that the uploaded medical records of the patient provide an extra layer of context that can help to inform the answers provided by ChatGPT Health.
“In terms of a balance between empowering patients and avoiding false confidence or misinformation, this is incredibly tricky. False information can easily arise, and a recent study suggested that some answers from chatbots in a test group of patients were no greater than using a search engine. However, part of this is often that the full context of the medical record is not available, or the values of the patient that a clinician could take time to understand.”
In a perfect world, users would treat ChatGPT as a supplemental service, not a trusted healthcare professional. That’s a huge expectation to have for millions of people using this platform every day, which is an interesting dilemma for the modern age. While it’s only another step in the evolution of AI, it’s a huge shift for patients and providers who will likely have to work together to navigate these murky waters.
AI in Healthcare: Promise and Peril
As platforms like ChatGPT Health expand the role of artificial intelligence in patient engagement, healthcare leaders are weighing both the opportunities and the risks. Here’s a quick look at where AI could help, and where caution is still warranted.
Three Potential Positives
1. Greater patient understanding
AI tools can help translate complex medical language into plain terms, allowing patients to better understand test results, diagnoses and treatment options between provider visits.
2. Improved access to information
With millions already turning to AI for health questions, these platforms offer on-demand guidance that can help patients prepare for appointments, navigate insurance decisions and track long-term health trends.
3. More empowered patients
By centralizing medical records and wellness data, AI may give patients a clearer picture of their health over time, encouraging more informed conversations with clinicians and stronger engagement in care decisions.
Three Potential Pitfalls
1. Risk of inaccurate information
AI systems can still produce incorrect or misleading responses. In healthcare contexts, even small errors could lead to confusion, delayed care or harmful decisions if users rely too heavily on AI guidance.
2. Privacy and data security concerns
Uploading sensitive medical records to third-party platforms raises questions about data ownership, confidentiality and how health information could be used beyond its original purpose.
3. Overreliance on technology
Experts caution that AI tools are meant to supplement – not replace – medical professionals. Patients who treat AI as a provider rather than a support tool may develop false confidence in automated advice.