Your editor touched it. Article 50's label exemption still might not apply.
A question I get more than almost any other on content labeling: “the draft was AI-written, but an editor went through it line by line before it published, so we don't need the label anymore, right?” The honest answer is: maybe, but probably not as often as teams assume when they ask it that way.
Article 50(4) does carve out a real exemption here. AI-generated text published to inform the public on matters of public interest needs a visible label, unless a person took substantive editorial responsibility for it. That exemption exists, and it's not a loophole, it's a deliberate part of the obligation. The part that trips teams up is what “substantive” is actually doing in that sentence.
It's doing a lot. A quick skim or a spell-check pass doesn't qualify. Neither does technically opening the document and clicking publish. The exemption is written for cases where someone genuinely took editorial responsibility for the piece, meaning it should be possible to point to who that person is and what they actually changed, not just assume it happened because a human was somewhere in the pipeline. “A human was involved” and “a human took editorial responsibility” are different claims, and only the second one is what the exemption describes.
Concretely, that's the difference between an editor who rewrote passages, checked claims, and would put their name behind the piece's accuracy, versus someone who opened the AI draft, read it once, and moved it to the CMS. Both look identical in a version-history diff that just shows “edited by a human account.” Only one of them is what Article 50(4) is actually asking for.
The part that's easy to miss entirely: this exemption is specifically about the visible label, Article 50(4). It doesn't touch the audit-trail obligation. Those are two different duties under Article 50, and getting an exemption on one doesn't get you out of the other. A piece that genuinely qualifies for the editorial exemption can skip the reader-facing “AI-generated” badge and still need a record that it started as AI output, if that's something your process wants to be able to show later, an editor, a legal team, or a future version of your own compliance review asking “wait, did we AI-draft this one?”
That distinction is also why Disclarion's SDK doesn't try to make this call for you. track_content() returns a fixed badge alongside the logged interaction, it doesn't know whether the text in front of it is about to go out untouched or get rewritten by an editor an hour later, and it shouldn't guess. Nothing about calling it forces you to render the badge it returns: you can log the interaction for your own record, the part Article 50's audit-trail duty actually wants, and still decide, as an editorial judgment call on your side, not to attach the visible label to a specific piece your team is confident earned the exemption. The SDK gives you the record either way. What you publish next to the content is a decision it deliberately leaves to you.
None of this is a reason to reach for the exemption by default. It's narrower than “a human looked at it,” it only covers one of Article 50's duties, and “we can probably argue this counted as substantive review” is a worse position to be in during an audit than a badge that costs you nothing to leave on. For the full breakdown of both content-labeling duties, see our content labeling guide, and for where this exemption sits relative to Article 50(2)'s separate machine-readable-marking requirement — a duty the editorial exemption doesn't touch at all — see our post on marking vs. labeling.
More from the blog
Your internal AI tools owe the same disclosure duty the customer-facing one does
An HR bot or IT helpdesk assistant used only by your own staff still interacts with natural persons. Article 50 doesn't carve out an exception for "internal-only."
Does the EU AI Act apply if you're not based in the EU?
"We're a US company, we don't have an EU entity" isn't the exemption it sounds like. The actual test is about who's on the other end of the conversation, not your mailing address.
Why I built Disclarion as an SDK, not a dashboard
The easier product to build is a dashboard you configure after the fact. I built Disclarion to sit inside the AI call itself instead, and it's worth explaining why.
Deepfakes get a stricter rule than an AI-written blog post
Article 50(4)'s deepfake duty isn't the same obligation as ordinary AI content labeling — it applies to image, audio, and video specifically, and the built-in exemptions work differently too.
Your chatbot isn't the only thing that owes a disclosure under Article 50.
Article 50's disclosure duty doesn't stop at chatbots — emotion recognition and biometric categorization systems carry the same obligation, and they're the two system types teams miss most often.
You named it "AI Assistant." That's not the same thing as disclosing it.
A bot's name feels like disclosure. Article 50 cares whether a reasonable user actually understood they were talking to an AI — not whether a label existed somewhere on the page.
You labeled the article. That's not the same thing as marking it.
Article 50(2)'s machine-readable marking and Article 50(4)'s visible label are two different duties. Images already meet the first one by accident. Article text usually doesn't meet it at all.
Your AI answers the phone now. There's no modal for that.
Article 50's disclosure duty doesn't go away because there's no screen. What actually discloses an AI voice agent on a phone call, when a modal isn't an option.
An AI audit trail doesn't need to know what your users actually said.
Teams hear "audit trail" and picture a database of every conversation. Article 50's version is narrower: it has to prove disclosure and labeling happened, not repeat what was said.
Article 50 skips disclosure when it's "obvious from context." That word is doing more work than it looks like.
Disclosure isn't required when AI use is obvious from context, but Article 50 never defines obvious, and most of the cases teams assume qualify, don't.
Your chatbot needs to disclose it's an AI. Once, not every message.
Disclosing a chatbot is AI sounds simple until you build it: repeat it every message and people tune it out, show it once and later visitors never see it.
You have logs. That's not the same thing as an audit trail.
Under Article 50, an audit trail has to be insert-only, complete, and producible on demand, not a data-engineering project when a regulator actually asks.
Your AI-generated content needs a label. Here's what that actually means.
A disclaimer someone remembers to add isn't a labeling system. Under Article 50, AI-generated content needs to be identifiable as such, reliably, not as a manual step someone can forget.
EU AI Act Article 50, explained: who it affects and what it requires
Article 50 of the EU AI Act has been enforceable since August 2, 2026: who it applies to, its three obligations, and the penalties for getting it wrong.