EU AI Act and chatbots — in force from 2 August 2026
Good news first: a customer-service or marketing chatbot is a limited-risk AI system under the EU AI Act. No licensing, no conformity assessment — the key obligation is transparency (Article 50). Here is what that means and what the ChatAutomata platform handles for you automatically.
What does Article 50 require?
Visitors must be able to tell they are interacting with an AI system — perceivably, within the conversation itself, at or before its start. A note buried in terms and conditions or a vague "assistant" name is not sufficient. The obligations apply from 2 August 2026; non-compliance can be fined up to €15 million or 3% of global annual turnover.
What ChatAutomata handles automatically
- Built-in AI label: every chat widget header permanently shows "AI assistant" — regardless of what you name your bot or how you configure it. This satisfies the in-conversation disclosure requirement.
- Human oversight: the operator takeover feature lets a real person join and take over any conversation at any time.
- GDPR package: privacy documentation, data-processing chapter on AI and automated decisions, data minimisation and EU data storage.
- No high-risk functions: the platform provides no biometric identification, emotion recognition or other Annex III functionality.
Your 5-point checklist as a deployer
- Do not present the bot as a human (e.g. a human name with a real photo and no AI indication) — the built-in AI label protects you, leave it visible.
- Mention the AI chatbot in your own privacy notice (our template chapter helps).
- If you publish AI-generated content (images, text) elsewhere, label it as AI-generated where the law requires.
- Make sure colleagues handling the chatbot understand its basic operation (AI literacy).
- Review your chatbot knowledge base regularly — accurate answers are the best compliance.
This page is general information, not legal advice. For specifics, consult your legal advisor.