Olivier Véran, former Minister of Health and founder of InnOV, a strategic consulting company specializing in tech and health topics, explains to JDN the challenges of the AI revolution in health.
JDN. Are doctors and patients already adopting AI? health ?
Olivier Véran. We are already seeing a very strong presence of AI in technical disciplines such as radiology or nuclear medicine, where it excels in detecting anomalies. In clinical medicine also, studies show that French doctors have already mostly used an LLM, such as ChatGPT, to support their diagnosis in a clinical case. Patients, too, massively use AI to obtain diagnostic or therapeutic advice. Faced with this reality, rather than denying it or crying danger, we must recognize that AI has already entered the practices of everyday medicine. Although there are limitations, such as bias, hallucinations, and the digital divide, that need to be addressed, AI is not about to leave the medical field any time soon.
But who is responsible in the event of a diagnostic error made by an AI?
On legal responsibility, the answer is clear: the doctor will always be legally responsible. Even if it uses validated and recommended AI, in the event of an error, the prescribing doctor will be responsible. The question of legal responsibility is therefore not up for debate. This is why I do not believe in AIs treating patients in total autonomy, without human supervision. The real question is rather that of the place of humans in an increasingly digital and AI-based medicine. Especially since certain studies show that, in specific cases, AI can be more efficient than the doctor. Sometimes doctors refuse to follow the AI’s recommendation, based on their experience. This could generate other legal questions, with patients who may ask why the AI’s advice was not followed.
Do you have concrete examples of the use of AI having an impact on mortality, diagnostic times or quality of life?
Several sectors are already showing tangible results. In the field of prevention, more and more French people are equipping themselves with connected watches or digital tools allowing them to monitor their physical activity and their health on a daily basis. Chatbots also allow patients better access to medical information. On the business side, solutions like the Ziwig Endotest have been validated by the HAS. The company has developed the first AI test based on salivary RNA to diagnose endometriosis, which affects 10% of women of childbearing age, the leading cause of infertility and whose diagnosis is delayed by an average of 7 years. AI also plays a growing role in oncology, in biotherapies and even in therapeutic compliance, by helping patients with chronic illnesses to better follow their treatments and improve their quality of life. Every year, hundreds of French companies innovate by sharing the ambition of better diagnosis and better treatment.
Announced in 2024, the France 2030 plan plans to mobilize 2.5 billion euros of investments in AI. Do you think the State is investing enough to transform the health sector with AI?
France is indeed investing significant amounts of public money in AI, part of which is dedicated to health. Of course, this remains incommensurate with the United States, where massive investment comes mainly from large technology companies. However, France has solid assets, including high-level research, innovative manufacturers, and some of the best mathematics schools in the world. The BPI also plays a key role in providing financial support to businesses. Now, the question will undoubtedly arise of choosing our battles, because we will not be able to be world leaders in all areas of AI with current budgets. We will need to identify strategic sectors where we already have excellent companies and concentrate our efforts to help them scale up and enable them to attract more private capital.
In a November 2025 report, the Court of Auditors estimated that the transformation of public action by AI remained “very disappointing”, given the ambitions displayed. How do you explain it?
There is still too little investment in AI within the State and its administrations. If we take the example of regional health agencies, their level of equipment is largely insufficient. Our administrations do not have the necessary skills internally to develop their own AI tools and must call on consulting firms. However, since the Covid crisis and criticism of this recourse deemed excessive, the latter can no longer intervene as necessary to implement AI in administrations. We therefore risk falling behind in this public sector compared to other European countries. If we do not quickly take the turn towards AI, while remaining vigilant about biases and hallucinations, we risk being downgraded in terms of public health. The benefits are real, particularly for better diagnosis and treatment, but they will not be measured in one year or in the short term in terms of return on investment.
How can we guarantee technological sovereignty on the subject of hosting health data, particularly in the face of Gafam?
With Mon Espace Santé, we have created a common platform allowing the interoperability of applications while offering guarantees of security, ethics and data confidentiality. The question of accommodation is a major issue on which work is underway. However, there is a certain ambivalence with some who request strictly French accommodation, while using tools like ChatGPT on a daily basis to obtain medical advice. In fact, the majority of applications used by French people today meet American or Asian standards.
More and more patients are turning to AI chatbots for information and even self-diagnosis. What is your view, as a doctor, on these practices?
It is impossible to prevent citizens from using digital tools to learn about their health, and this reappropriation of the body and health is rather positive. Anything that helps to acculturate the French and give them more autonomy seems beneficial to me. This could undoubtedly also help avoid certain health and understanding problems, as we saw during the Covid crisis. On the other hand, it is essential that this advice be validated by health professionals. An AI cannot replace a doctor and cannot provide prescriptions in any case. But by encouraging screening in the event of symptoms or by providing a better understanding of diseases, certain uses of these tools seem useful to me.
In conclusion, what is your vision of the role that AI will play in the medical field in the coming years?
AI will permeate all health sectors, as digital technology has done before. Healing without AI will become extremely rare. However, it will remain a medical assistant, intended to amplify the professional’s diagnostic and decision-making abilities. AI will extend the doctor’s action, including in the patient’s home, via chatbots or software dedicated to chronic diseases. Unlike IT, which has sometimes made hospital work heavier by adding administrative tasks, AI will make it possible to simplify and streamline the practices of professionals and provide immediate responses. We will be able to diagnose diseases earlier and treat them more effectively, while improving the management of chronic pathologies and cancer. We are therefore at the start of a revolution in the health field, which is not ready to stop.
Olivier Véran is a neurologist. He was a deputy and general rapporteur of the social security budget before being appointed Minister of Solidarity and Health, then Minister of Democracy and government spokesperson. He founded a strategic consulting company, InnOV, specializing in tech and health issues.
