EU AI Act for Marketing & AdTech: Transparency, Deepfakes & Art. 50
Marketing runs on generative AI — synthetic imagery, chatbots, personalised copy and, increasingly, deepfakes. Under the EU AI Act (Regulation (EU) 2024/1689), most of this sits in the limited-risk tier governed by Art. 50 transparency duties rather than the high-risk regime. But 'limited' does not mean unregulated: disclosure and labelling are mandatory, and manipulative techniques can be outright prohibited.
Is it in scope?
Marketing and AdTech AI is generally not listed in Annex III, so it usually escapes the high-risk tier. Instead it is caught by Art. 50 transparency: chatbots must disclose they are AI, AI-generated or manipulated audio, image, video and text must be marked as artificially generated, and deepfakes must be labelled. These duties apply from 2 August 2026. Separately, Art. 5 prohibits AI that uses subliminal, manipulative or deceptive techniques to materially distort behaviour and cause harm — a real constraint on aggressive persuasion and targeting tactics.
Typical AI use cases
- Generative ad copy and creative production
- AI-generated product and campaign imagery
- Conversational marketing and sales chatbots
- Synthetic voice and video (deepfake) ads
- AI-driven audience targeting and personalisation
- Automated social and email content generation
Risk classification
Most AdTech AI is limited-risk, and the marketing company deploying a generative model is typically its provider for disclosure purposes. Your obligations are practical: label AI-generated and synthetic content, disclose chatbot interactions, and clearly mark deepfakes under Art. 50 (from 2 August 2026). The harder line is Art. 5: manipulative or deceptive systems designed to distort consumer behaviour to their detriment are prohibited outright — so persuasion techniques, not just disclosure, need review. High-risk only enters if AI is repurposed for an Annex III use.
Obligations to prepare for
FAQ
Do we have to label AI-generated ad images and copy?
Yes. Art. 50 requires providers to mark AI-generated or manipulated audio, image, video and text as artificially generated in a machine-readable way, with deepfakes clearly labelled. This applies from 2 August 2026.
Is targeted advertising with AI prohibited?
Not as such. Personalised targeting is generally limited-risk, but Art. 5 prohibits subliminal, manipulative or deceptive techniques that materially distort behaviour and cause harm — so the method matters, not just the disclosure.
Does our marketing chatbot need a disclaimer?
Yes. Under Art. 50 you must inform users they are interacting with an AI system unless it is already obvious from the context. The requirement applies from 2 August 2026.
Run the free 2-minute risk checker — no sign-up.