Until now, AI in healthcare has been limited to indirect applications: assisting doctors with diagnoses, analyzing medical images, sorting medical records, and optimizing care pathways. With ChatGPT Health, OpenAI is ushering in a much more profound shift. For the first time on a large scale, a conversational AI is designed to interact directly with patients’ personal medical data and provide them with an intelligible interpretation of it.
This shift is significant. It marks the transition from AI that supports the healthcare system to AI that engages directly with patients, aiming to help them understand their own bodies, test results, treatments, and risks. A technological—but above all cultural and ethical—threshold has just been crossed.
When AI Becomes the Interpreter of Medical Records
One of the major limitations of modern medicine stems from the growing complexity of the data it generates. Lab reports, imaging reports, genetic test results, medical histories, prescriptions… Medical records have become so extensive that they are often incomprehensible even to the patient.
ChatGPT Health aims precisely to capture this complexity. The AI does more than simply aggregate data; it contextualizes it, rephrases it, and connects the various pieces of information. An abnormal lab result is no longer just an isolated number, but a piece of information viewed within the context of the patient’s medical history, age, current treatments, and known risk factors.
This interpretive capability is based on models trained on vast biomedical and clinical datasets, capable of linking symptoms, test results, and probable diagnoses. According to OpenAI, the goal is not to make a diagnosis, but to enhance the patient’s understanding and preparedness for the medical consultation1.
Practical applications that are reshaping the patient-caregiver relationship
In practice, ChatGPT Health is part of a series of very concrete scenarios that reflect a quiet but profound shift in everyday medical practice.
For example, a patient with a chronic condition can obtain a clear overview of how their key health metrics have changed over several months, identify periods of instability, and better understand the impact of their treatments. For patients taking multiple medications, AI helps clarify potential drug interactions and organize the questions they should ask their doctor.
As part of preventive care, ChatGPT Health also helps connect scattered data—such as family history, lifestyle habits, and lab results—to identify subtle warning signs that are often overlooked due to a lack of time orclarity.²
These practices are part of a broader trend: according to a McKinsey study, nearly 60% of patients in developed countries report that they do not fully understand their medical documents, a lack of understanding that is directly linked to lower treatment adherence3.
Key figures: a market under pressure and undergoing significant restructuring
The emergence of ChatGPT Health comes amid rapid growth in medical AI. The global market for artificial intelligence applied to healthcare is estimated to exceed $150 billion by 2030, with an annual growth rate of over 35%4.
At the same time, pressure on healthcare systems is mounting. According to the OECD, demographic aging could lead to a 20–25% increase in demand for healthcare by 2035, without a corresponding increase in the number of healthcare workers5. AI is thus no longer seen as a technological luxury, but as a tool for managing complexity.
However, only 32% of patients currently say they trust AI to handle their sensitive health data, a figure that highlights the gap between technological potential and social acceptance6.
Access ChatGPT: Availability, Cost, and Schedule
Unlike ChatGPT’s general-purpose features, ChatGPT Health is being rolled out cautiously and under strict supervision. The solution is currently being piloted in the United States in partnership with healthcare providers that comply with FHIR interoperability standards and HIPAA regulatory requirements7.
Access is provided directly through the ChatGPT interface, without a separate app, but requires explicit and granular consent from the patient for each connected data source. OpenAI states that medical data is neither stored for advertising purposes nor used to train general-purpose models.
From an economic standpoint, ChatGPT Health is positioned as a premium offering. Initial industry estimates suggest a subscription price ranging from $30 to $60 per month, depending on the level of analysis, data volume, and activated features8.
In Europe, the timeline remains uncertain. The requirements of the GDPR and the entry into force of the AI Act impose additional safeguards. OpenAI is working with European partners, but a large-scale rollout before 2026 seems unlikely, given the significant compliance challenges involved.
ChatGPT Health and MedGPT: Two Approaches to Medical AI
The rise of healthcare-focused chatbots is not limited to initiatives by major U.S. players. Alongside ChatGPT Health, a French solution like MedGPT illustrates a different approach to medical AI, one that places greater emphasis on data sovereignty and compliance with the European regulatory framework. Comparing these two tools provides a better understanding of the technological, ethical, and strategic choices that currently shape the development of AI in healthcare.
| Criteria | ChatGPT Health (OpenAI) | MedGPT (France) |
| Positioning | A general-purpose AI assistant enhanced with healthcare features, integrated into the OpenAI ecosystem | AI assistant specialized in healthcare, designed from the ground up for medical use |
| Origin and Governance | An international company based in the United States | A French initiative within a European framework |
| Data used | A combination of general medical knowledge and contextual data provided by the user | Specialized medical corpora, primarily French-language and European |
| Link to medical records | Gradual access to health data, based on the user's preferences and permissions | A more cautious approach, often limited to the analysis of manually entered data |
| Key use cases | Understanding symptoms, medical education, decision support, personalized health monitoring | Medical interpretation assistance, patient information, support for healthcare professionals |
| Target audience | The general public, patients, and non-specialist users | Patients, healthcare professionals, stakeholders in the healthcare system |
| Regulatory framework | Announced compliance with international standards, gradual adaptation to local regulations | Strong alignment with the GDPR and European health data requirements |
| Major ethical issues | Large-scale management of sensitive data, model trust and transparency | Data sovereignty, clinical reliability, medical liability |
| Strategic ambition | Making ChatGPT a comprehensive and personalized health companion | Offer a sovereign, specialized alternative in medical AI |
A particularly sensitive ethical boundary
While ChatGPT Health is fascinating, it also raises concerns. Giving an AI access to personal medical data means encroaching on one of the most sensitive aspects of privacy. The issue is no longer just one of technical security, but of cognitive delegation.
AI that explains, prioritizes, and contextualizes can also shape perceptions of risk, influence decisions, or create a false sense of control. Several researchers warn of the danger of an algorithmic medicalization of everyday life, in which AI becomes a constant voice in the interpretation of the body9.
International institutions emphasize the need to uphold a clear principle: AI does not replace medical diagnosis or clinical decision-making, but serves as a tool for understanding and mediation10. Without this framework, there is a risk of shifting medical responsibility to opaque systems.
Toward more accessible medical care—but with conditions
ChatGPT Health holds great promise: to make healthcare more accessible, more seamless, and more personalized. It could help reduce the information gap between patients and healthcare providers, improve treatment adherence, and strengthen preventive care.
But this promise will only be fulfilled under strict conditions: transparency in the models, data governance, informed consent, and clear integration into the care pathway. Otherwise, medical AI risks becoming yet another opaque intermediary, precisely where it claims to bring clarity.
Learn more
This breakthrough by OpenAI in the healthcare field is part of a broader landscape of medical AI that is rapidly taking shape. In a complementary vein, we invite you to read our analysis “MedGPT: The Free French Medical AI That Rivals ChatGPT”, which explores the emergence of specialized solutions designed to meet the clinical, regulatory, and ethical requirements specific to the healthcare sector. The article highlights issues such as data sovereignty, the reliability of recommendations, and professionals’ trust in these new intelligent medical assistants.
References
1. Topol, E. (2023). Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again. MIT Press.
https://mitpress.mit.edu
2. Nature Medicine. (2024). AI-assisted patient data interpretation. https://www.nature.com/nm
3. McKinsey Global Institute. (2025). The economic potential of generative AI in healthcare.
https://www.mckinsey.com
4. PwC. (2024). Global AI in Healthcare Market Outlook.
https://www.pwc.com
5. OECD. (2024). Health workforce challenges and digital transformation.
https://www.oecd.org
6. Edelman. (2024). Trust Barometer: Health and AI.
https://www.edelman.com /a>
7. U.S. Department of Health and Human Services. (2024). HIPAA and AI-enabled health tools.
https://www.hhs.gov
8. CB Insights. (2025). AI health startups and pricing models.
https://www.cbinsights.com
9. Stanford HAI. (2024). Human-centered AI in healthcare.
https://hai.stanford.edu
10. World Health Organization. (2023). Ethics and governance of artificial intelligence for health.
https://www.who.int/a>
