From 2 August 2026 the EU can enforce AI Act transparency rules against any business running a chatbot or publishing AI content that doesn't disclose it isn't human.
Why It Matters
This obligation, Article 50, applies to any company that runs a website chatbot or voicebot, or publishes AI-generated text, images, or voice, not just to AI model builders like OpenAI or Anthropic. Enforcement power over these rules activates on this date. For an SME the fine cap is lower than the headline figure, but it still reaches up to 3% of global annual turnover.
What To Do About It
Check every AI touchpoint a customer meets: website chatbot, automated email replies, AI-generated marketing copy or voiceovers. A visible line such as "You're chatting with an AI assistant" in the chat window satisfies the chatbot requirement; label AI-generated content the same way. Do it now, enforcement power is live from this date.
Sources
Related Signals
The EU AI Act's next enforcement wave lands on 2 August: GPAI penalty powers and Article 50 transparency duties go live,
22 Jul 2026
German news reports that US authorities can access corporate data stored on servers in Frankfurt, despite their physical location in Europe.
29 Mar 2026
An EU Parliament committee has endorsed a proposal to delay the full enforcement of the AI Act.
25 Mar 2026