Your AI answers the phone now. There's no modal for that.
Every disclosure pattern I've written about so far assumes a screen. A modal on the first message, a small badge that stays visible for the rest of the session, a caption on an image. All of that falls apart the moment the interaction is a phone call, and AI voice agents answering support and sales calls are common enough now that this isn't a hypothetical. Article 50 doesn't carve out an exception for “there was nowhere to put a modal.”
For the obligation itself, I covered it in a longer explainer on Article 50 a while back: people need to be told they're interacting with an AI system, unless it's obvious from context. Nothing about that requirement is specific to visual interfaces. It just happens that most of the tooling, including the disclosure patterns I've written about before, was designed around chat widgets, where a screen is a given. Voice removes that assumption without removing the obligation.
The good news is the underlying logic doesn't actually change, only the medium does. The modal-once-then-badge pattern works because it says something clearly at the start, once, instead of repeating it into background noise. The audio equivalent is the same shape: a short, spoken disclosure at the start of the call, before the substance of the conversation begins, not buried after a hold message or mentioned only if the caller happens to ask. “You're speaking with an automated assistant” before the call gets into why the person called is doing the same job a modal does on a chat widget's first message.
The part that doesn't translate is the persistent badge. There's no audio equivalent of a small icon that just sits in the corner for the rest of the session. In practice this matters less than it seems: a phone call is already a single, bounded session in a way a chat window isn't, it doesn't sit open and forgotten in a browser tab for three days. Said clearly once, at the point someone actually picks up, covers the interaction it's attached to. What doesn't work is saying it only once, at the very start of an IVR menu, and then transferring the caller through two more automated steps that never mention it again.
The obvious-from-context exception applies here too, and it's actually a cleaner fit for voice than it is for chat. A phone menu that says “press 1 to talk to our AI assistant” before routing the call has arguably already disclosed it, the same way a page that's explicitly a model demo has. A voice agent designed to sound indistinguishable from a human rep, with no announcement anywhere in the call flow, is the harder case, and it's the same “obvious to the team, not obvious to the caller” problem I wrote about before, just with a voice instead of a chat bubble.
None of this changes what has to be logged, either. An audit trail doesn't care whether the interaction happened over text or audio, it needs the same facts either way: which provider handled it, when, and whether disclosure was actually given for that specific call. A voice agent calling Disclarion's SDK logs the interaction the same way a chat integration does, because the record was never built around what the interaction looked like on screen. The SDK doesn't generate the spoken line for you, that part is on your call flow, but it doesn't need to: the disclosure is metadata your system acts on, not text baked into a UI component, which is exactly what makes it work for a channel that has no UI at all.
Voice won't be the last channel this comes up for, either. Whatever the next interface is that doesn't look like a chat window, the question is the same one worth asking now: does “we disclose AI use” describe something a person on the other end would actually notice, or just something that's true on the channels where it was easy to add.
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.
Your editor touched it. Article 50's label exemption still might not apply.
Article 50(4) waives the visible AI-content label when a human takes substantive editorial responsibility for a piece — but a quick skim doesn't qualify, and the exemption doesn't touch the audit-trail duty underneath it.
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.
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.