If your chatbot pretends to be a person, that is now an enforcement matter in the EU
The EU has begun enforcing the AI Act rule that customer-facing bots must say they are bots. The fix costs a sentence. The fine is up to €15 million.
What changed
The EU's AI Office has begun enforcing Article 50 of the AI Act — the transparency provision. In plain terms: if a person is talking to an AI system, they must be told they are talking to an AI system. Reported penalties for getting this wrong run to €15 million or 3% of global turnover, whichever hurts more.
This is the part of the AI Act that touches ordinary businesses first. Not the exotic provisions about biometric surveillance — the chatbot on your website.
Why it matters to you
If you sell to anyone in the EU and there's an assistant on your site, in your booking flow or answering your support email, this applies to you regardless of where your business sits. And plenty of off-the-shelf chat widgets are configured — sometimes by default — to present as a friendly human with a stock photo and a first name.
The deeper point is about trust rather than fines. A customer who discovers mid-conversation that "Sarah" was software doesn't just feel misled about the chat; they wonder what else you're dressing up.
What we'd do about it
Ten-minute audit: open every customer-facing channel you run and ask whether a reasonable person would know they're talking to software. If the answer is no, add the disclosure — one sentence at the start of the conversation. We put one on every AI-drafted article on this site for the same reason, and we've yet to see any evidence that honesty reduces engagement. If your chat vendor makes disclosure hard to configure, that tells you something about the vendor.
What to watch
Enforcement appetite. First actions tend to target the egregious cases, but the paper trail you want — a note saying you reviewed your channels and when — costs an hour now and is very hard to reconstruct later.
Sources
Every claim in this article links to a primary source. The agent collects them; the editor verifies each one before publication.
- AI Agent News — Week of August 10, 2026 — AI Agent Store , 2026-08-10
- EU AI Act, Article 50 (transparency obligations) — EU AI Act , 2026
How this was made
Drafted by our research agent from the sources above, then edited, fact-checked and approved by a human before publishing. We publish this note on every article because we think you should know.
More insights
AI agents that browse the open web are already hitting closed doors
Sites are starting to block autonomous agents. If your plan depends on an AI acting like a junior on the public internet, treat that as a fragile assumption.
Meta’s new business terms cover your website, not just your posts
From 30 October, Meta’s commercial terms for business accounts reportedly require AI training on linked website content, with no opt-out except leaving the platforms. Treat that as a licensing and data-protection review, not a footnote.
Square puts voice ordering on a kiosk
Block’s payments firm is selling a 21.5-inch ordering kiosk that talks back. For a shop that already has a counter, that is a staffing question, not a gadget story.