Guides · Article 50 · EU AI Act
What healthcare AI teams should know about Article 50
Healthcare-adjacent products ship AI that talks to patients — symptom checkers, intake assistants, wellbeing chatbots — and, in some cases, systems that infer emotional state. Article 50 of the EU AI Act sets transparency duties for both. Read this caveat first: healthcare is an area where AI systems may engage far more of the Act than its transparency rules. Article 50 covers only that slice; depending on your systems, other parts of the EU AI Act — including the high-risk-system requirements — may also apply, alongside other EU law for health products. Confirm your full obligations with qualified counsel. This guide is informational only, not legal advice.
Patient-facing chatbots and assistants — Article 50(1)
Article 50(1) provides that providers must ensure AI systems intended to interact directly with natural persons are designed so those people are informed they are interacting with an AI system, unless that is obvious to a reasonably well-informed, observant and circumspect person in context (Regulation (EU) 2024/1689). In a health setting the question of who is answering — a clinician, a support agent, or software — is rarely trivial to the person asking, which makes the 'obvious' exception a cautious bet. Article 50(5) points toward a clear, distinguishable disclosure at the latest at the first interaction, meeting applicable accessibility requirements — a point with extra weight where users may rely on assistive technology.
Emotion-inferring systems — Article 50(3)
Where a deployed system infers emotion or categorises people from biometric data — for example in patient monitoring or telehealth analytics — Article 50(3) requires deployers to inform the natural persons exposed of the operation of the system and to process personal data in accordance with the EU data-protection instruments the paragraph cites. Whether a given tool is an 'emotion recognition system' under the Regulation is fact-specific, and in health contexts the data-protection overlay deserves its own review with your data-protection adviser.
What healthcare teams commonly document
- Each patient- or user-facing AI system and where it interacts directly with people.
- The AI-interaction disclosure wording, placement, and accessibility notes.
- For emotion-inferring systems: the notice to exposed persons and the data-protection cross-references.
- A counsel-reviewed assessment of which parts of the Act beyond Article 50 — including high-risk requirements — and which other health-sector rules apply.
Common questions
Does a symptom-checker chatbot need an AI disclosure?
Article 50(1) addresses AI systems intended to interact directly with people, subject to the 'obvious' exception. Whether it applies to your checker is fact-specific — and healthcare AI may separately engage high-risk requirements elsewhere in the Act and other EU rules for health products. Confirm the full picture with qualified counsel.
See what may apply to your business
Answer seven quick questions for an automated, informational indication of which Article 50 obligations appear likely to apply — free, and not legal advice.