What specific disclosure does a clinical AI chat interface need to make to comply with Article 50?

The system must be designed so that any person interacting with it is clearly informed they are interacting with an AI rather than a human, through an explicit and understandable disclosure rather than an ambiguous or buried notice. For digital therapeutics or clinical decision support tools with conversational interfaces, this typically means a clear statement at the start of or throughout the interaction, not just a mention in terms of service.

How does the synthetic content labeling requirement apply to MedTech patient education materials?

If a device or digital health product generates synthetic audio, image, video, or text, such as an AI-generated patient education video or voice narration, that output must be marked as artificially generated or manipulated in a machine-readable format under Article 50. Manufacturers producing AI-generated patient-facing content need a technical labeling mechanism embedded in the content itself, not simply a disclaimer elsewhere in the product, to satisfy the machine-readable requirement.

What additional obligation applies if a system performs emotion recognition or biometric categorization?

Article 50 requires deployers of emotion-recognition or biometric-categorization systems to inform the persons exposed to that system, which is a distinct and additional obligation from the general AI-interaction disclosure. MedTech tools that incorporate affect or emotion detection as part of a monitoring or diagnostic function should evaluate whether this triggers the specific exposed-persons notification requirement in addition to the standard AI disclosure obligation.

All terms

RegulatoryRegulated Pathways AI / ML in Devices

AI Act Transparency Obligations

Article 50 requirements for specific AI systems, including disclosure that a user is interacting with AI and labelling of synthetic content.

Reviewed by Christian Espinosa, Founder, Blue Goat CyberLast reviewed July 25, 2026

Definition

Article 50 requires providers to design AI systems that interact with natural persons so that those persons are informed they are interacting with an AI, and that synthetic audio, image, video, or text output is marked as artificially generated or manipulated in a machine-readable format. Deployers of emotion-recognition or biometric-categorisation systems must inform exposed persons.

What the regulation says

Regulation (EU) 2024/1689, adopted on 13 June 2024, entered into force on 1 August 2024. Article 50 applies from 2 August 2026. Under Article 50(1), providers must design systems that interact directly with natural persons so those persons know they are interacting with AI, unless this is obvious to a reasonably well-informed, observant and circumspect person in the circumstances. Article 50(2) requires providers of systems generating synthetic audio, images, video or text to make outputs machine-readable and detectable as artificially generated or manipulated. Technical solutions must be effective, interoperable, robust and reliable as far as technically feasible; specified exceptions include certain standard editing assistance. For MedTech deployments, Article 50(3) requires deployers of emotion-recognition or biometric-categorisation systems to inform exposed persons and comply with applicable data-protection law. Article 50(4) separately requires deployers to disclose AI-generated or manipulated deepfakes and certain AI-generated public-interest text, subject to specified exceptions and qualifications. These disclosure duties are distinct from providers’ machine-readable marking duties. Article 50(5) requires the relevant information to be clear and distinguishable, provided no later than first interaction or exposure, and compliant with applicable accessibility requirements. Transparency compliance does not itself establish medical-device conformity or authorise otherwise prohibited AI uses.

What this means in practice

Digital therapeutics and clinical chat interfaces should surface an explicit AI disclosure and, where they generate synthetic content (for example patient education videos), meet the labelling requirement. This applies even when the device is not high-risk.

Examples

A MedTech application providing personalized health coaching via a chatbot explicitly states at the beginning of each session: 'You are now interacting with an AI-powered health assistant.'
A digital therapeutic that generates animated videos explaining medical conditions includes a persistent, machine-readable watermark and text overlay: 'This video contains artificially generated content.'
A clinical decision support tool that uses AI to summarize patient records clearly displays a banner at the top of the summary: 'AI-generated summary, always verify with original sources.'
An AI-powered diagnostic imaging analysis tool includes a disclaimer with its results: 'AI analysis performed. Review and confirm results with a qualified healthcare professional.'
A virtual reality-based therapy system that uses AI to adapt scenarios informs the patient at startup: 'Your experience is being dynamically adjusted by an AI system.'
A MedTech platform offering AI-generated patient education leaflets includes a footer on each document: 'This content was created using artificial intelligence.'

Common pitfalls

•Failing to implement clear, user-facing disclosures that explicitly state interaction with an AI system is occurring. This can lead to user confusion and non-compliance, even if the AI is not deemed high-risk.
•Not adequately labeling synthetic content in a machine-readable and easily understandable format. Simply adding a small, obscure disclaimer may not meet the intent of informing users about artificially generated material.
•Assuming that transparency obligations only apply to high-risk AI systems. Many transparency requirements, like informing users of AI interaction and labeling synthetic content, apply more broadly regardless of risk classification.
•Overlooking the specific requirements for deployers of emotion-recognition or biometric-categorisation systems to inform exposed persons. This is a distinct obligation beyond general AI interaction disclosures.

Frequently asked questions

Article 50's transparency obligations apply regardless of whether the AI system is classified as high-risk, meaning even a non-high-risk digital health tool or chat interface that interacts with natural persons must comply. This is a distinct obligation layer from the high-risk system requirements in Articles 8 through 15, so a device team cannot assume that because their system escaped high-risk classification, it has no AI Act transparency duties at all.

Grouped by theme

Editor's picks

· Hand-selected related concepts

Regulatory

High-Risk AI System (EU AI Act)

AI systems captured by Article 6 of the EU AI Act because they are safety components of regulated products or listed in Annex III.

Market Segments

Digital Therapeutics(DTx)

Software-based interventions that prevent, manage, or treat disease.

More in Regulatory

· Same category

Regulatory

510(k) Premarket Notification(510(k))

FDA submission demonstrating a device is substantially equivalent to a legally marketed predicate.

Regulatory

510(k) Summary vs 510(k) Statement

Two alternative disclosure mechanisms in a 510(k) submission, a Summary is public on FDA's website; a Statement promises to share Safety & Effectiveness info on request.

Regulatory

513(g) Request for Information(513(g))

Formal mechanism to ask FDA whether a product is a device and, if so, its likely classification.

Regulatory

Abbreviated 510(k)

510(k) variant that relies on FDA guidance, special controls, or recognized consensus standards.

Cited by

Where this term appears across MedTech Terms.

Ecosystems (2)

Latest in MedTech

Sources

3 sources

Every citation below opens the original document. Each is graded against our source-tier hierarchy so you can see what rests on binding law versus commentary.

Tier 1Binding law and standards· 1Tier 2Regulator guidance and consensus· 2

Link health: 3 verified· last checked 2026-06-20

EUR-Lex·1FDA·1European Commission·1

Inline markers like [1] jump to the matching reference above.

My progress

All terms0 / 513

Regulatory0 / 101

Saved on this device, and to your account when signed in. View studied terms

Full glossary

Browse all terms A to Z, each with a short definition.

On this term

Category: Regulatory
Sources: 3
Updated: 7/25/2026

Compare with another term

Learn in 60 seconds

Article 50 requirements for specific AI systems, including disclosure that a user is interacting with AI and labelling of synthetic content.

·This applies even when the device is not high-risk.
·Deployers of emotion-recognition or biometric-categorisation systems must inform exposed persons.

Remember this

Watch out: Failing to implement clear, user-facing disclosures that explicitly state interaction with an AI system is occurring. This can lead to user confusion and non-compliance, even if the AI is not deemed high-risk.

Related terms

You may also need

Auto-suggested from Regulatory and shared keywords.

All Regulatory terms

From the Blue Goat network

Related resources and services on this topic.