Europe's AI Disclosure Law Went Live, and Most Products Complied by Adding One Sentence

The EU's transparency obligation under the AI Act — requiring chatbots, voice agents and generated media to disclose that a human is not on the other end — moved from law-on-paper to active enforcement in early August. Regulators had spent over a year warning companies this date was coming, and it showed: most consumer-facing AI products complied by inserting a short disclosure line into onboarding flows or first-message templates rather than redesigning anything structural. The harder compliance question isn't the front-of-house chatbot, where a one-line disclosure satisfies the letter of the rule. It's the AI-assisted layer buried inside otherwise-human workflows — a support agent whose replies are AI-drafted and lightly edited, a sales call where a voice model handles the first few minutes before handing off. Regulators have signaled that disclosure obligations extend to these blended cases, but enforcement guidance on exactly where the line sits is still thin, and companies operating in that gray zone are making their own judgment calls for now. For businesses building AI-facing products for the EU market, the practical lesson is that disclosure requirements are becoming a checklist item like a cookie banner, not a design constraint — provided the AI involvement is real and visible somewhere in the flow, not concealed behind a human-sounding persona.