Chatbot Disclosure 101 — The 5-Minute Fix You Need by August 2

If your website has a chatbot, you need to add one sentence by August 2, 2026.

That's it. One sentence.

This is possibly the easiest AI Act requirement to knock out. But ignore it, and you won't meet Article 50 disclosure requirements when they kick in. Let's fix it right now.

What the requirement actually says

Article 50 of the EU AI Act says users must know they're talking to AI "no later than the first interaction."

Translation: before someone types their first message, they need to see something that says "this is AI."

That's the whole thing. No consent checkbox. No legal disclaimers. Just clarity.

Who this affects

If you have any of these, this applies to you:

Website chatbots (support, sales, lead gen)
Automated messaging on your product (in-app assistants)
Customer support AI (even if it escalates to humans later)
Email autoresponders powered by AI
Any interface where a user thinks they might be talking to a person, but they're not

Estimated impact: 40%+ of SMBs use some form of chatbot. If that's you, this is relevant.

5 disclosure texts you can copy-paste right now

Here's what meets the requirement. Pick one, adapt it, ship it.

Option 1 — Direct and simple:

"This chat is powered by AI. Responses are automated."

Option 2 — Friendly:

"You're chatting with our AI assistant. Need a human? Just ask."

Option 3 — Branded:

"Meet Ada, our AI support bot. She's here 24/7 to help."

Option 4 — Explicit handoff:

"This is an AI assistant. If you'd prefer to speak with our team, type 'human.'"

Option 5 — Ultra-minimal:

"AI Assistant"

All of these work. The law doesn't specify tone or length. It just says users can't be deceived. Make it obvious. Make it visible. You're done.

Where to place it

Three options, depending on your setup:

Option A — Chat widget header The best spot. Put "AI Assistant" or your disclosure text in the header of the chat window. Visible before the first message.

Option B — First automated message If your chat opens with a greeting like "Hi! How can I help you today?", add your disclosure there:

"Hi! I'm an AI assistant. How can I help you today?"

Option C — Visual indicator Some tools let you add a small badge or icon. A robot icon + "AI" label next to the chat bubble counts.

Technical implementation notes

If you're using Intercom, Drift, Zendesk, or similar: most platforms added AI disclosure settings in 2025-2026. Check your settings panel under "AI" or "Automation." Many have a one-click disclosure toggle now.

If you built your own: add a <div> with the disclosure text above your chat input. Style it so it's not hidden by default. If you're using a framework like Vercel AI SDK or LangChain, inject the disclosure into the system prompt's first visible message.

If you're not technical: copy this email template below and send it to whoever manages your website.

Email template for your team

Subject: Quick AI Act fix needed — chatbot disclosure

Hey [Name],

We need to add a disclosure to our website chatbot by August 2, 2026 (EU AI Act requirement). This is a 5-minute fix.

What to do:
Add this text to our chatbot, visible before the first user message:
"This chat is powered by AI. Responses are automated."

Where to put it:
- Chat widget header, OR
- First automated greeting message, OR
- Visual label next to the chat icon

Most platforms (Intercom, Drift, etc.) have a setting for this now. If not, just add the text manually.

Let me know when it's live? Thanks.

[Your name]

Forward that. Done.

The "obvious AI" exception

There's a gray area: if it's extremely obvious that the interaction is automated, you might not need a disclosure.

Examples where it's obvious:

A chat widget that says "Frequently Asked Questions" and serves canned responses
A form with dropdown menus labeled "AI Search"
A tool explicitly branded as "AI Assistant" in giant letters

But here's the thing: adding one sentence takes 5 minutes. Why risk it? Just add the disclosure. The cost is near zero, and the clarity is worth it.

This is about transparency, not permission

Important: you don't need user consent to use a chatbot. You just need to be clear about what it is.

You're not asking permission. You're providing context.

The AI Act's transparency requirements are designed to reduce confusion, not ban automation. You can keep your chatbot. You can keep your AI support. Just tell people what they're talking to.

Why this matters (beyond the legal requirement)

Honestly? This is good UX.

When users know they're talking to AI, they adjust their expectations. They don't get frustrated waiting for a "human" response. They phrase questions more clearly. They understand limitations.

Transparency builds trust. It doesn't break it.

The actual deadline

Article 50 transparency obligations take effect August 2, 2026. That's the date general-purpose AI systems and certain transparency rules kick in.

You've got time. But this fix is so fast, why wait?

One more thing

If you're using AI in other places — email automation, content generation tools shown to users, AI-generated images on your site — check if disclosure is needed there too. Article 50 covers more than just chatbots.

Same principle: if a user might not realize they're interacting with AI, tell them.

Your 5-minute action plan

Copy one of the 5 disclosure texts above
Log into your chat platform (or email your web team)
Add the disclosure to your chat widget header or first message
Preview it to confirm it's visible before the first interaction
Done

That's it. You're ready.

Copy our chatbot disclosure template and implement today

Need help figuring out what else the AI Act requires? Run a free scan at Act-Ready — we'll tell you exactly what applies to your company and what doesn't.

This document supports readiness preparation. It does not constitute legal advice.