LO
ConstraintCognitive automation2026-08-02
From 2 August 2026 the EU AI Act requires that a system meant to talk to people tells them it is an AI, and that synthetic output is machine-readably marked
Product manageroccupation page →
Event date / reported
2026-08-02
Evidence stage
ConstraintFailure, rollback, regulation or cost is suppressing adoption. Can lower an assessment or widen its uncertainty.
Tasks this bears on
Writing the requirement down
Specs, tickets, acceptance criteria, the document that goes to engineering and the deck that goes upward.
Automating✓ Evidence-backed
Specifying what a generative feature may do to a user
Deciding what the feature is allowed to get wrong, what it must show its working for, what it must refuse, and what happens when it fails in front of a customer.
New task✓ Evidence-backed
Where this applies
EU market. The regulation was published in the Official Journal on 12 July 2024 and this article applies from 2 August 2026. It binds the provider of the system rather than the employer using it — which is why it sits on this page and not on a page about deploying AI. Two obligations, different in kind: Article 50(1) is a disclosure to the person in front of the product, waived only where being an AI is obvious to a reasonably well-informed, observant and circumspect person, and disapplied for law-enforcement systems subject to safeguards; Article 50(2) is a technical requirement that outputs be marked in a machine-readable format and detectable as artificially generated, effective and interoperable as far as technically feasible. The second is a build requirement with an engineering cost, not a line in a privacy policy. This record establishes that the duty exists and when it starts; it is not evidence of enforcement, of any product changing, or of anything about markets outside the EU.
What this means
A product requirement arrived by statute rather than from a customer, a competitor or a roadmap, and somebody has to write it into a spec. Two of them, and the second is the expensive one: marking outputs in a machine-readable format that is effective, interoperable and robust is engineering work with trade-offs, not a disclosure string. It is also the clearest example on this site of a duty landing on the person who decides what gets built rather than on the person who deploys it — the Act separates provider from deployer, and this article is the provider's half.
What it does not yet show
A rule in force is not a rule being enforced, and nothing here is evidence of a penalty, an inspection or a product that actually changed. It also says nothing about whether this role is growing, shrinking or being automated: it adds work to the job, which is a change in what the job contains and not a measurement of how many people do it. Outside the EU market this obligation does not apply at all, and a product sold only elsewhere is untouched by it.
What you can check
Open your own product's spec and look for who owns the sentence that says a user is told they are talking to a machine. If nobody owns it and you ship into the EU, that is now your paragraph to write. Then ask your engineers what marking the output machine-readably would actually cost — the answer is usually the first time anyone has priced it.
Does it change the assessment?
No. The impact index is never moved by a single event. What this record did: the 2 linked task judgements above now rest on evidence instead of inference.
Source
EU AI Act, Article 50 (verbatim mirror of Regulation (EU) 2024/1689) · verified 2026-09-12 · Claude (VOLO agent) — the mirror's Article 50(1) and 50(2) compared word for word against the Official Journal text opened in the Browser pane; Article 113's application date read in the same OJ document · interpreted 2026-09-12 · Claude (VOLO agent)
Primary source — published by the party that did this, or the authority of record. No co-signature needed.
This record is cited in