EU AI Act Article 50 Checker

How to use

1. Question 1: say honestly whether your AI system’s output could be used in the EU — a public website or app open to EU visitors counts as “possibly”.
2. Question 2: place yourself as provider (the system is on the market under your name), deployer (you use someone else’s system in your business), both, or not sure. Questions 3–4 then map the interaction and content duties.
3. Read your result tier and the next steps it generates, copy the result link to share it, and check the key-dates table below for when each duty lands.

About this tool

Two things decide most real cases. The first is role: a provider places a system on the market under its own name or trademark (Article 3(3)), while a deployer uses a system under its own authority (Article 3(4)) — and branding a vendor’s system as your own can move you from the second column into the first. The second is reach: under Article 2(1), the Act covers third-country providers and deployers where the output of the system is used in the Union, which is why a US company with an EU-facing website cannot wave the question away. Article 50 has applied since 2 August 2026; the only transitional relief is narrow — providers whose generative systems were already on the EU market before that date have until 2 December 2026 for the machine-readable marking duty alone (Article 111(4), inserted by the Digital Omnibus, Regulation (EU) 2026/1744).

In keeping with how this site works, the checker runs entirely in your browser: your answers are never sent to a server — they exist only on this page and, if you use it, in the result link you copy. It covers Article 50 only. The rest of the AI Act — prohibited practices, high-risk duties, general-purpose AI model duties — and other EU law such as the GDPR are separate questions, and nothing here is a substitute for an EU AI Act counsel reviewing your actual product.

How it works

Article 50 is a list, not a principle: four duty blocks, each with its own actor and trigger. (1) Interaction — providers ensure that AI systems intended to interact directly with natural persons are designed so those persons are informed they are interacting with an AI system, unless this is obvious to a reasonably well-informed, observant and circumspect person in context (Art. 50(1)). (2) Marking — providers of AI systems, including general-purpose AI systems, generating synthetic audio, image, video or text content ensure outputs are marked in a machine-readable format and detectable as artificially generated or manipulated, with technical solutions that are effective, interoperable, robust and reliable as far as technically feasible (Art. 50(2)); the duty does not apply to the extent the system performs an assistive function for standard editing or does not substantially alter the deployer’s input or its semantics. (3) Exposure — deployers of an emotion recognition system (Art. 3(39)) or a biometric categorisation system (Art. 3(40)) inform the natural persons exposed of the operation of the system and process personal data under the GDPR (Art. 50(3)). (4) Publication — deployers disclose deep fakes, and AI-generated or manipulated text published to inform the public on matters of public interest (Art. 50(4)). Whatever the block, the information must be provided in a clear and distinguishable manner at the latest at the time of the first interaction or exposure, and must meet applicable accessibility requirements (Art. 50(5)).

Role decides which blocks are yours. “Provider” (Art. 3(3)) means a person or body that develops an AI system or a general-purpose AI model, or has one developed, and places it on the market or puts the system into service under its own name or trademark, paid or free. “Deployer” (Art. 3(4)) means a person or body using an AI system under its authority, except in a personal non-professional activity. The blocks follow the definition: 50(1) and 50(2) bind providers; 50(3) and 50(4) bind deployers. The same company is routinely both — provider of its own product, deployer of a vendor’s chatbot — and the duties stack feature by feature. The definition, not the contract’s label, controls: putting your own name or trademark on a system someone else built is the classic way a “reseller” becomes a provider. Reach is the third leg: Article 2(1) applies the Regulation to providers placing systems on the Union market, to deployers in the Union, and to third-country providers and deployers where the output produced by the AI system is used in the Union.

Timing and consequences. The Regulation applies from 2 August 2026 (Art. 113), and Article 50 — in Chapter IV — is not among the provisions given a different date, so the transparency duties have applied since 2 August 2026. One transitional rule exists: providers of AI systems generating synthetic audio, image, video or text content that were placed on the market before 2 August 2026 “shall take the necessary steps in order to comply with Article 50(2) by 2 December 2026” (Art. 111(4), inserted by Regulation (EU) 2026/1744, the Digital Omnibus on AI). It covers the marking duty only, for providers only, for pre-August systems only; paragraphs 1, 3, 4 and 5 were not deferred. Non-compliance with the transparency obligations in Article 50 is subject to administrative fines of up to EUR 15,000,000 or, for an undertaking, up to 3% of total worldwide annual turnover, whichever is higher (Art. 99(4)(g)); for SMEs the lower amount applies (Art. 99(6)). The Commission’s Article 50 guidelines (adopted 20 July 2026) and the voluntary Code of Practice on Transparency of AI-Generated Content (published 10 June 2026) are guidance, not law — but they are the reference implementation regulators will compare against.

Worked examples

US SaaS sells its own AI writing assistant: A Delaware company sells a subscription AI writing tool under its own brand; EU customers sign up on the public website. It has the system developed and places it on the market under its own name — a provider under Article 3(3), in reach under Article 2(1). If the tool chats with users, Article 50(1) interaction disclosure is a design duty; its generated text output falls under the Article 50(2) machine-readable marking duty. Result pattern: likely provider obligations.

Online shop embeds a vendor’s support chatbot: A retailer adds a well-known vendor’s chat widget to its store, including its EU storefronts. The retailer did not develop the system and does not market it under its own name — it uses it under its own authority, a deployer under Article 3(4). The 50(1) design duty sits with the vendor, but the retailer should confirm the disclosure exists, keep it visible, and never skin the bot to look like a human agent. Result pattern: likely deployer obligations, with vendor-check next steps.

News site publishes AI-drafted articles: A publisher uses an AI tool to draft news articles for the public. That is AI text published to inform the public on matters of public interest, squarely in Article 50(4) deployer territory: disclose it — unless each piece goes through genuine human review or editorial control and a named editor holds editorial responsibility for publication. A spell-check pass does not meet that bar under the Commission’s guidelines. Result pattern: likely deployer obligations, driven by the public-interest text answer.

Internal analytics tool, no EU-facing features: A company uses AI internally to summarise its own support tickets: no customer-facing interaction, nothing published, no content generation for the public, no emotion or biometric features. Even though the business has EU customers for other products, this system triggers none of the four Article 50 situations. Result pattern: likely out of scope for Article 50 — re-checked when a customer-facing feature ships.

Quick reference

Frequently asked questions

More free tools