Everyone fixated on “high-risk” AI missed the rule that touches almost everyone.

Article 50 of the EU AI Act — the transparency obligations — becomes enforceable on August 2, 2026. Unlike the high-risk requirements, it does not care whether your system is high-risk. It cares whether a human is interacting with AI without knowing it.

Three things become mandatory: disclosing when a person is interacting with an AI system, labeling AI-generated or manipulated content in a machine-readable way, and identifying deepfakes. If you run a customer-facing chatbot, generate marketing content with AI, or use synthetic media anywhere in your funnel, this is your rule — not someone else’s.

The high-risk obligations affect the few. Article 50 affects the many. And it is the one most organizations have not even mapped.

The number that matters

Zero. That is how many of your AI customer touchpoints are exempt from Article 50 simply because they are “low-risk.” Risk tier does not grant an exemption from disclosure. A friendly support bot is still a system a customer is talking to without being told.

Build an AI touchpoint map this week

List every place AI speaks to, or produces content for, a customer or the public:

  • Chatbots and virtual assistants
  • AI-generated email, ad copy, and social content
  • Synthetic voice or video
  • Recommendation and personalization surfaces

For each, answer one question: would a reasonable person know they are dealing with AI? If the answer is no, you have a disclosure gap — and the fix is usually a sentence, not an engineering project. The cost of doing it now is a label. The cost of doing it later is an enforcement action.

How LANStatus helps

We help you build the touchpoint map, flag where disclosure is missing, and put the simple, compliant fixes in place before the deadline — so a rule almost nobody is reading does not become the reason a regulator comes knocking.

Walk your own funnel as a customer. How many times does AI talk to you without telling you?

We map your customer-facing AI and close the disclosure gaps before the deadline. Let's talk.

Explore Professional Services

Brian Diamond

Founder & CEO, LANStatus · Fractional Chief AI Officer

Brian founded LANStatus in 2001 and works with mid-market healthcare and financial-services organizations on AI strategy, governance, and security. He publishes The CAIO Brief, a weekly briefing for leaders navigating AI in real time.

The CAIO Brief · BrianOnAI · LinkedIn

A version of this article first appeared in The CAIO Brief.