A general practitioner using an algorithm to analyze a chest X-ray must now adhere to a precise ethical framework. Since a decree published in the Official Journal on July 29, 2026, a new article R.4127-13-1 of the code of ethics requires practitioners using artificial intelligence or telemedicine to ensure the safety of care and the confidentiality of data.
Medical news is no longer limited to laboratory discoveries; it also involves the regulatory obligations that govern their concrete deployment.
Ethical obligation and medical AI: what the July 2026 decree changes
Before this text, no article in the code of ethics explicitly mentioned artificial intelligence. Decree No. 2026-691 of July 27, 2026, fills this gap. It creates a formal ethical obligation for any physician using AI, telemedicine, or other technical innovations.
In practice, the practitioner must ensure that the digital tool they use complies with the recommendations of health authorities and the Order Council. We are talking about verifiable precautions, not a vague principle of caution.
This evolution shifts health AI from a purely technical realm to an individual responsibility of the physician. A radiologist relying on a nodule detection algorithm takes on ethical responsibility for the reliability of this tool. Specialized publications on health innovations rarely cover this regulatory angle, and it is regular monitoring of medical news that helps not to miss out, especially through portals like https://doctinews.fr/ that relay this type of development.
Digital health compliance certificate: the access lock to funding
Since December 31, 2024, every digital health service must obtain a compliance certificate issued by the Digital Health Agency. This certificate covers three areas: interoperability, security, and ethics. Without it, it is impossible to feed the Shared Medical Record, obtain Ségur listing, or access public funding (SONS, France 2030, ARS).
For medical software publishers, this constraint changes the game. An uncertified EHR (electronic health record) simply cannot be deployed in an establishment that wishes to benefit from Ségur funding. It is observed on the ground that several publishers have had to accelerate their updates to stay competitive.
Portability of health data: the law of May 2026
Law No. 2026-403 of May 26, 2026, adds an additional layer. It creates a legal obligation for the portability of health data. A patient must be able to retrieve their data from a digital service and transfer it to another, in a usable format.
This provision aims to prevent lock-in to a proprietary ecosystem. For medical-social establishments or city practices, this means that the choice of software should no longer be made solely based on features, but also on the publisher’s ability to guarantee this portability.
Ségur of digital health: where are the concrete deployments
The Ségur digital program has been structuring the modernization of health information systems in France for several years. Wave 2 of Ségur extends the scope to medical-social establishments, with dedicated funding for children and child protection.
Feedback varies on this point depending on the regions and types of structures. Some ESSMS (social and medical-social establishments and services) have already listed their UDR (computerized user record), while others struggle to identify a certified publisher that meets their specific needs.
- The Ségur wave 2 listing now concerns EHRs and UDRs in the health, medical-social, and child sectors, with funding directed by the ARS.
- Publishers like Docaposte Santé have had their two EHRs listed, completing the secure data sharing chain.
- Speech therapists and other paramedical professions are gradually being integrated into the scope of digital Ségur, which expands the network of shared data.
This gradual deployment illustrates a often underestimated point: innovation in health is as much about the interoperability of systems as it is about fundamental research.
Health data and prevention: an underutilized use
The health data collected by digital devices and connected objects represent a resource for prevention. We can identify weak signals in the data of diabetic patients, anticipate cardiac decompensations, or detect epidemiological trends at the territorial level.
The reality on the ground is more nuanced. Prevention is advancing, but not always where expected. The establishments that truly exploit this data for preventive purposes remain in the minority. The barriers are multiple: heterogeneous quality of collected data, lack of team training, absence of standardized protocols to transform an algorithmic alert into clinical action.
AI and medical devices: the framework of the AI Act
The gradual entry into force of the European regulation on artificial intelligence (AI Act) classifies AI software used as medical devices in the high-risk category. This classification imposes requirements for transparency, traceability, and human oversight.
For a French publisher, this means documenting the entire lifecycle of the algorithm: training data, performance metrics, bias management. An algorithm not compliant with the AI Act will not be able to obtain CE marking as a medical device, which blocks its market entry in Europe.
Governance of AI in healthcare establishments
The question is no longer whether hospitals will use AI, but how they will organize it. The governance of artificial intelligence in healthcare establishments requires concrete choices: who validates the deployment of an algorithm, who monitors its performance over time, who manages incidents.
Few establishments have today formalized an AI governance policy. Hospital IT departments often deal with tools adopted service by service, without a consolidated vision. Field feedback shows that the structures making the most progress are those that have appointed an AI referent or created a dedicated committee involving clinicians, IT professionals, and legal experts.
The medical news in the coming months will likely be marked by the publication of institutional recommendations on this subject. Technological innovations in health are plentiful, but their adoption increasingly depends on the ability of establishments to structure their digital governance and to train their teams in the new regulatory frameworks.
