Does My Chatbot Need to Disclose It’s AI Under EU Law Right Now?
Your product has an AI chatbot. It talks to customers, answers questions, maybe closes deals.
Did you know it might already be violating the law?
As of this month, the answer could be yes, and the fine on the table is not small. If your users include anyone in the EU, the rules changed on August 2, 2026, whether you noticed or not.
The Deadline That Already Passed
On August 2, 2026, the EU AI Act’s transparency rules under Article 50 became enforceable. Any AI chatbot or AI-generated content reaching individuals in the EU must disclose that it is AI, at the first interaction, every time. Penalties run up to 15 million euros or 3 percent of global turnover, whichever is higher.
Location of Your Users Matters, Not Your Address
This rule does not care where your company is incorporated. It turns on whether your AI output reaches someone located in the EU. If your chatbot talks to EU users, you are in scope, even without an EU office or an EU contract.
Your Chatbot’s Promises Can Now Bind You in Court
In May 2026, Germany’s Higher Regional Court of Hamm ruled a company was legally bound by what its own AI chatbot told a customer. Courts are treating chatbot statements the same as statements from an employee. That means:
- a discount your bot invents becomes a real obligation
- a policy your bot misstates can become the policy a court enforces
- an unauthorized promise from your AI agent can still cost you money
Common Founder Mistakes
- Treating the Whole AI Act as Delayed. Founders hear “AI Act delayed” once and stop tracking it. That word covers two different tracks: (a) high-risk system rules, pushed to December 2027 or (b) transparency and disclosure rules, live and enforceable since August 2, 2026. Ignoring the difference means missing a deadline that already passed.
- Assuming “We Don’t Sell Into the EU” Clears Us. Many founders check their sales contracts and billing addresses and conclude they are safe. The rule is triggered by AI output reaching individuals located in the EU, not by where your company sells or invoices. A free trial user browsing from Berlin is enough.
- Burying an AI Disclaimer in the Terms of Use. Two gaps show up in almost every product review: (a) a disclaimer buried in the Terms of Use instead of shown at the moment of interaction and (b) no written record of what the AI agent is actually authorized to promise a customer. Courts are already exploiting both.
10-Minute Self-Check
Before you push another product update, work through this:
- Does my chatbot tell the user it is AI the moment the conversation starts?
- Does that disclosure repeat, not just appear once buried in a legal page?
- Do I know whether any of my users are located in the EU right now?
- Have I written down what my AI agent is and is not authorized to promise?
- Has anyone on my team reviewed our chatbot scripts since August 2, 2026?
- Would our disclosure hold up if a regulator or a customer’s lawyer looked at it today?
If you cannot answer yes to all of these, you are not ready to call your product compliant yet.
Bottom Line
The EU AI Act’s transparency rules are not a future problem. They are enforceable today, with fines that scale to global turnover. A chatbot that talks without disclosing itself, and without a clear record of what it is allowed to promise, is a liability sitting inside a product that otherwise looks ready to launch.
Is My AI Chatbot Actually Compliant Before My Next Launch?
Our launch-ready legal package is tailored to your software, your customers, and the way your product actually operates. Schedule a free 30-minute discovery call to discuss your business, your goals, and whether our team can help prepare your product for launch.
Book here: https://calendly.com/primumlaw/30min