Skip to main content
Connic
Back to BlogIndustry Insights

EU AI Act Article 50: What Your AI Agent Must Disclose

The Commission adopted its final Article 50 guidelines on 20 July 2026, thirteen days before the rules apply. Here is what agent teams have to disclose, and when.

July 26, 202610 min read

On 20 July 2026 the European Commission adopted its final guidelines on the transparency obligations in Article 50 of the EU AI Act. Those obligations start applying on 2 August 2026, thirteen days later. If you run an agent that talks to people, generates content, or publishes text, Article 50 hands you at least one of five disclosure duties, and the guidelines are the first detailed account of what satisfying them looks like.

The practical questions are which trigger applies to which kind of agent, where the exceptions are thinner than they look, and which parts of the work your own application has to do rather than the platform underneath it. For the wider risk-tier picture and the high-risk timeline, read our guide to EU AI Act compliance for AI agents.

What the Commission Published on 20 July

The package explains how the Commission reads Article 50 without changing a word of it. Alongside the guidelines the Commission issued a Communication approving their content, a public FAQ on the Article 50 obligations, and a short fact sheet. The goal is to get the same rules read the same way in every member state, so that people can tell when they are dealing with an AI system and when content was generated or altered by one.

None of it is binding. Only the Court of Justice can authoritatively interpret the Regulation. But the guidelines are the reference point that national market surveillance authorities and the AI Office will work from, so they are the best available preview of how those authorities will read the text from August onward.

The final text also confirms that the Commission assessed the Code of Practice on Transparency of AI-Generated Content as adequate. The Code was published by the AI Office on 10 June 2026 and adherence is voluntary. Signing it is a way to demonstrate compliance with the marking duties, not a formal legal presumption that you have met them. Skip it and nothing is automatically wrong, but you carry the full burden of showing that whatever marking you chose is good enough.

The Five Disclosure Triggers

Article 50 carries five separate obligations, split between the provider role and the deployer role, and a single system can trigger several of them at once.

Direct interaction, Article 50(1), provider
A system intended to interact directly with natural persons must be designed so those people are informed they are dealing with an AI. The Commission FAQ names chatbots and AI agents as the central examples.
Synthetic content marking, Article 50(2), provider
Providers of systems generating synthetic audio, image, video or text must mark the output in a machine-readable format and make it detectable as artificially generated. The duty is qualified by technical feasibility, implementation cost, and the acknowledged state of the art.
Emotion and biometric systems, Article 50(3), deployer
Deployers of emotion recognition or biometric categorisation systems must inform the people exposed to them.
Deepfakes, Article 50(4), deployer
Deployers publishing image, audio or video deepfakes must disclose that the content is artificially generated or manipulated. Three tests all have to hold: the content closely resembles real people, objects or places; the subject exists or could plausibly have existed; and the result would falsely appear authentic to a viewer.
Public-interest text, Article 50(4), deployer
AI-generated or manipulated text published to inform the public on matters of public interest must be labelled. The Commission lists politics, public administration, justice, fundamental rights, security, health, environment and consumer safety among the qualifying topics.

Article 50(5) governs how all of it is delivered: clearly and distinguishably, at the latest at the first interaction or exposure, and in a form that meets accessibility requirements. For deepfakes the guidelines are explicit that a machine-readable mark alone is not enough. A person has to be able to perceive the disclosure.

The Trigger That Catches Most Agent Teams

The guidelines set four cumulative criteria for whether a system interacts directly with a natural person. All four have to hold:

  • 1.The system qualifies as an AI system under the Regulation
  • 2.There is a genuine two-way exchange, not merely collection of data from the person
  • 3.The AI communicates with the person itself, rather than through a human intermediary
  • 4.The counterpart is a natural person

That draws a usable line through a typical agent fleet. An agent answering customers over the Telegram connector, replying from a shared inbox through the email connector, or streaming into a support widget over a WebSocket session is squarely in scope. An agent that consumes a Kafka topic, enriches records, and writes to Postgres is not: there is no natural person on the other end of the exchange. Background systems and machine-to-machine communication are excluded.

The duty attaches to the provider, meaning whoever develops the system and places it on the market or puts it into service under their own name. If you build an agent and offer it to your customers, that is you. Because the obligation attaches at design time, the disclosure has to be built into the system rather than remembered at runtime.

Map one agent to its Article 50 triggers

Bring a customer-facing agent and we can walk through which disclosure duties it triggers, what the transparency record needs to contain, and which controls Connic covers.

Talk through your Article 50 scope

The Exceptions Are Narrower Than They Look

Obviousness has a legal test behind it

Article 50(1) does not apply where it is obvious to the person that they are dealing with an AI. The guidelines measure that against a hypothetical average person who is reasonably well-informed, observant and circumspect, and they interpret the carve-out restrictively given how central transparency is to the Regulation. Being confident internally that everyone knows the support widget is a bot is a different thing from testing it against that standard and writing down the result. An authority will ask for the written version.

Editorial control means someone actually read it

The public-interest text obligation falls away where the content underwent human review or editorial control. The guidelines define that as a deliberate examination of the substance by one or more natural persons with relevant knowledge, plus a person or organisation who can approve, alter or reject the content on substantive grounds and who holds editorial responsibility for publication. Running a spellchecker over the output does not qualify, and neither does a cursory glance.

For teams shipping content agents this is the exception most likely to matter, and the one that turns into an evidence problem. Relying on it means being able to show who reviewed which output, when, and with what authority to reject it.

Marking has real carve-outs, and they are technical

Article 50(2) does not reach systems performing an assistive function that does not substantially alter the input data. The guidelines add further exclusions for short sequences of symbols, source code, machine-to-machine outputs, and closed-loop development environments. Evidently artistic, satirical or fictional works keep a reduced deepfake obligation: the disclosure must exist, but in a manner that does not hamper the display or enjoyment of the work.

The Dates That Matter

10 June 2026
The AI Office publishes the Code of Practice on Transparency of AI-Generated Content. Voluntary adherence, assessed as adequate in the final guidelines.
20 July 2026
The Commission adopts the final Article 50 guidelines, with a Communication, an FAQ, and a fact sheet.
2 August 2026
Article 50 applies. Systems launched on or after this date carry the full set of duties from day one, with no run-in period.
2 December 2026
End of the transition for generative systems already on the market before 2 August. It covers the Article 50(2) machine-readable marking duty only, not the interaction or labelling duties.

Retroactivity splits by modality, and the split is easy to get backwards. Images, audio and video generated or manipulated before 2 August 2026 do not have to be marked retroactively. Text works the other way. Text generated before 2 August but published on or after that date does need labelling, unless the editorial control exception applies. Scheduled content queues and evergreen libraries are worth checking on that basis.

What non-compliance costs
Breaching Article 50 falls under Article 99(4) of Regulation (EU) 2024/1689: administrative fines up to 15 million euros or 3% of total worldwide annual turnover for the preceding financial year, whichever is higher. For SMEs, Article 99(6) applies whichever of those figures is lower.

Who Marks the Output When You Bring Your Own Key

Article 50(2) sits with the provider of the generative system, and the technical expectations behind it live where content is produced: cryptographically signed provenance metadata, imperceptible watermarks that survive format conversion, and fingerprint registries as a supplement. The Code of Practice is explicit that no single technique currently satisfies all four quality criteria it sets out, which are effectiveness, interoperability, robustness and reliability, so it pushes providers toward layered approaches rather than one mechanism.

That changes how you read your own obligations on Connic. Connic is bring-your-own-key and never hosts or resells models, so machine-readable marking of model output is a property of the model provider you choose and of the application you put in front of users. It is not something an agent runtime can retrofit onto content it did not generate. What you can do is capture the upstream provider's marking as evidence supporting your own record, which is why the fifth item in the checklist below is a question for your model provider. Their evidence supports your record. It does not transfer your responsibility.

Turning Article 50 Into an Operational Record

Disclosure duties fail audits in a predictable way. The banner exists, the label is there, and nobody can produce a dated record showing which system carries which duty, on what basis, and that the disclosure reached the person it was meant for. Connic AI Governance, an Enterprise entitlement, is where that record lives. It is documentation and evidence rather than enforcement, and it is scoped per project. Read the AI Governance reference for the full workflow.

Assessments that pin the trigger
An assessment is an immutable, versioned answer set covering EU scope, operator roles, risk routes and the Article 50 dimensions. Connic derives a preliminary result from a pinned catalog version and records the regulatory source it was reviewed against. Recording new answers creates a new version rather than overwriting the old one.
One transparency policy per system
The policy documents how your organisation handles each trigger: direct interaction, synthetic content, deepfakes, public-interest text, and biometric or emotion use. It records the disclosure copy and its locales, the placement, the display mode, and whether the disclosure fires at first interaction or on every response.
Attestations as delivery evidence
Transparency applications record that a disclosure was applied, failed, or could not be confirmed, each with an attesting party, timestamp, evidence reference and factual basis. Provenance is either a customer application attestation or an upstream provider attestation, which is where the model provider's marking evidence lands.
Exports auditors can verify
An evidence snapshot is an immutable, metadata-only capture of systems, assessments, controls, policies, applications, monitoring plans and incidents, plus run, approval and judge telemetry. It is content-addressed with a SHA-256 hash and streams as an archive with per-file integrity hashes. Raw prompts and model outputs are deliberately excluded.
Where the boundary sits
Connic stores the policy and the supporting evidence. It does not modify prompts, responses, streams or connector traffic, does not inject prefixes or banners, and does not watermark media. Your application implements and delivers every disclosure. The assessment result is preliminary and is not a legal determination.

Approvals are the editorial-control paper trail

If you plan to rely on the human review exception for public-interest text, the guidelines require substantive examination by a person who can reject the content. That is exactly what an approval gate in front of publication produces: the output, the named reviewer, the decision, and the timestamp. Pair it with the trace behind each run and you can show the exception applies instead of asserting that it does.

Where to Start Before 2 August

  • 1.List every agent whose output reaches a natural person, and note the surface it reaches them through. That list is your Article 50(1) scope
  • 2.For each one, decide your role. Building and offering the agent generally makes you a provider, using someone else's system under your own authority makes you a deployer, and both can apply at once
  • 3.Write the disclosure copy, in every locale you serve, and place it so it lands at or before the first exchange rather than in a settings page or a footer
  • 4.If you publish AI-generated text on public-interest topics, choose deliberately between labelling it and running genuine editorial review, then build the evidence trail for whichever you picked
  • 5.Ask your model provider what marking they apply to generated media and how it can be detected, and record their answer as upstream evidence

For how Connic's controls line up against the wider Regulation, read the EU AI Act platform statement. None of this is legal advice, and a preliminary assessment is not a classification: the final call on which duties you carry belongs to you and your counsel.

Frequently Asked Questions

Article 50 transparency obligations apply from 2 August 2026. The European Commission adopted its final implementation guidelines on 20 July 2026. Generative AI systems already placed on the market before 2 August 2026 have until 2 December 2026 to meet the machine-readable marking duty in Article 50(2), but that transition does not extend to the interaction, deepfake, or public-interest text duties.

If the agent is intended to interact directly with natural persons, yes. The Commission FAQ names chatbots and AI agents as central examples of Article 50(1). Four criteria must all hold: the system qualifies as an AI system, there is a genuine two-way exchange rather than mere data collection, the AI communicates with the person itself rather than through a human intermediary, and the counterpart is a natural person. Background systems and machine-to-machine communication fall outside it.

Disclosure is not required where it is obvious to the person that they are interacting with an AI system. The guidelines assess this against a hypothetical average person who is reasonably well-informed, observant and circumspect, and interpret the exception restrictively. An internal assumption that users know the system is a bot is not equivalent to a documented assessment against that standard.

The Article 50(2) machine-readable marking duty sits with the provider of the generative AI system, and the technical mechanisms live where the content is produced: signed provenance metadata, imperceptible watermarking, and fingerprint registries. On a bring-your-own-key platform such as Connic, which never hosts models, marking is a property of the model provider you select and the application you ship. Upstream provider evidence can support your compliance record without transferring your responsibility.

The guidelines require deliberate examination of the substance of the content by one or more natural persons with relevant knowledge, together with a person or organisation who can approve, alter or reject it on substantive grounds and holds editorial responsibility for publication. Spellchecking or a cursory review does not qualify. Teams relying on this exception need a record of who reviewed which output, when, and with what authority.

Under Article 99(4) of Regulation (EU) 2024/1689, non-compliance with the Article 50 transparency obligations carries administrative fines of up to 15 million euros or up to 3% of total worldwide annual turnover for the preceding financial year, whichever is higher. Article 99(6) applies whichever figure is lower for small and medium-sized enterprises.

More from the Blog

Industry Insights

The OpenAI Hugging Face Hack: Guardrail Lessons for AI Agents

OpenAI models escaped a test sandbox and breached Hugging Face in July 2026. What the incident reveals about guardrails and how to secure production AI agents.

July 24, 20269 min read
Industry Insights

Soofi S Preview: Access, Benchmarks & AI Agent Fit

Soofi S is a gated preview of Germany's 31.6B open-model project. Its benchmarks, release status, limits, and potential for self-hosted AI agents.

July 15, 20269 min read
Industry Insights

AI Agent Platforms With EU Data Residency: 2026 Shortlist

A 2026 shortlist of AI agent platforms grouped by EU residency model, including coverage for traces, storage, and model calls.

July 6, 202612 min read
Industry Insights

Connector Patterns for AI Agents: Webhook vs Kafka vs Postgres vs SQS

Webhook, Kafka, Postgres LISTEN/NOTIFY, and SQS each trigger an AI agent differently. Compare their delivery guarantees, ordering, and durability to pick one.

June 29, 20269 min read
Industry Insights

State of AI Agents in DACH 2026

How DACH teams build, trigger, and run production AI agents in 2026: adoption, model mix, connectors, cost, reliability, and compliance, from Connic customer data.

June 27, 202612 min read
Industry Insights

Pre-built Connectors for AI Agents: Skip the Integration Glue

Production agents have to receive events from and send results to the systems around them. Pre-built connectors turn that plumbing into a platform feature instead of custom code you write and maintain.

June 16, 20268 min read
Industry Insights

How to Run AI Agents in the EU Without US Hyperscalers

Run production AI agents in the EU without US hyperscalers: what EU-hosted must really mean, where the US CLOUD Act exposes you, and a sovereignty checklist.

June 4, 20269 min read
Industry Insights

AI Agent Deployment Platforms in 2026: 4 Types Compared

Compare agent-native runtimes, workflow engines, framework stacks, and plain frameworks on lock-in, pricing, and connectors.

April 19, 202612 min read
Industry Insights

The EU AI Act Is Here. Your AI Agents Need to Comply.

The EU AI Act is changing how teams deploy AI agents. Learn which obligations apply and how Connic makes approvals, audit trails, guardrails, and observability part of the platform.

April 13, 202611 min read