One searches for a symptom on their phone between two appointments, stumbles upon a forum, then a video from an anonymous account recommending to stop a treatment. Three minutes later, one hesitates to call their doctor or follow the advice of a stranger. This common scenario illustrates the real problem: it is not the access to health information that is lacking online, but the ability to sort what is reliable from what is not.
Automated summaries and AI-generated responses: an additional layer of risk
Since the rollout of Google’s AI previews in France on July 22, 2026, health search results no longer only link to websites. The engine can assemble a summary from heterogeneous sources: a news article, a brand page, a post on a social network.
The problem is real. One reads a synthetic answer at the top of the page, takes it for medical consensus, while it sometimes mixes institutional sources with promotional or unverified content. Checking the sources cited by the AI response has become as necessary as verifying the site itself.
On Portail Santé, content is categorized by theme and linked to identified professionals or organizations, allowing for quick cross-referencing of what has been read in an automated summary.
Identifying the author and conflicts of interest on a health page
The French law of June 9, 2023, on commercial influence, supplemented by texts adopted in 2024, 2025, and 2026, imposes enhanced reporting of commercial partnerships and limits claims about health products. In practice, many pages still do not comply with these obligations or apply them discreetly.
When encountering health content online, three checks take less than a minute and filter out the majority of dubious sources:
- Is the author named, with verifiable qualifications (doctor, pharmacist, researcher affiliated with a recognized organization)? An article signed by an anonymous “editorial team” without a scientific committee deserves more caution.
- Does the page display a publication or update date? Medical information older than three years may be outdated, particularly regarding treatments and HAS recommendations.
- Is there a commercial link between the author and a product mentioned in the content? The presence of affiliate links, promo codes, or purchase buttons should raise alerts, even if the content appears objective.
These criteria do not guarantee reliability on their own, but they eliminate a large portion of problematic content.
Institutional sources and secure spaces: two distinct levels of reliability
People often confuse two things. On one hand, public institutional sites (Santé.fr, HAS, Inserm) publish validated content accessible without registration. On the other hand, Mon espace santé offers access to personalized medical documents in a secure environment, with encrypted messaging and shared medical records.
Both are useful, but not for the same needs. To understand a pathology or verify the side effects of a medication, institutional sites and databases like Vidal remain the reference. To retrieve an examination report or communicate with one’s treating physician without going through a regular email, Mon espace santé serves a different role.
Mon espace santé: a deployed but underused tool
According to the report from Assurance Maladie cited by Selectra in September 2026, the service is widely deployed but not effectively used by its beneficiaries. The secure messaging and document submission features remain unknown to part of the population, even though they directly address a need for confidentiality that open research does not cover.
Health influencer content: what the law now regulates
Social media has become a major channel for health information, especially for those under 35. The problem is not that the information is always false, but that we do not always have the tools to assess its quality.
Since the law of June 9, 2023, and its supplements, influencers must visibly disclose commercial partnerships. Claims about health products or services are regulated. In practice, compliance varies across platforms and content creators.
To filter health content on social media, one can apply the same criteria as for a website: identified author, verifiable qualifications, cited sources, absence of undeclared commercial links. An account that mixes promotion of dietary supplements and medical advice without mentioning a partnership should be excluded from one’s usual sources.
Building a verification routine suited to one’s needs
Rather than memorizing a list of trusted sites, one saves time by adopting a systematic reflex. When reading health information online, cross-referencing with at least one institutional source (HAS, Santé.fr, Vidal, Inserm) rarely takes more than two minutes and changes the quality of the decision that follows.
For common situations (mild symptoms, vaccination reminders, side effects), public resources cover the majority of questions. For complex situations or treatment decisions, no online source replaces a medical consultation. Online research prepares the discussion with the doctor; it does not replace it.
Inserm has also launched an educational kit aimed at middle and high school students to analyze false health information, a sign that source verification is becoming a basic skill, on par with reading a medication leaflet.
