The EU Artificial Intelligence Act, the world’s first comprehensive legal framework for AI, is now in force. If you’re a marketer operating in, or targeting consumers within, the European Union, compliance is no longer optional. This guide breaks down everything you need to know: who it applies to, what’s banned, and how transparency rules affect your campaigns.

Who does the EU AI Act apply to?

The regulation covers the entire AI value chain. It applies to:

  • Providers: any company or individual that develops an AI system and places it on the market under their own name or trademark, whether based in the EU or abroad, and whether the system is free or paid.
  • Deployers: any person or organisation using an AI system under their own authority within the Union. Individuals using AI solely for personal, non-professional purposes are exempt.
  • Third-country operators: providers and deployers based outside the EU, if the output of their AI system is intended for use inside the Union.
  • Importers, distributors, and manufacturers: those bringing AI systems into the EU market or integrating AI into products.

Practical implication for marketers: if your AI-generated campaigns, ads, or content reach EU consumers, you are within scope — regardless of where your company is headquartered.

Provider vs. deployer: what’s the difference?

Most marketing teams will be deployers — using AI tools built by providers (e.g. image generators, copywriting tools, chatbot platforms) to create content and interact with customers. Providers face heavier technical and documentation obligations; deployers face operational and transparency requirements.

RoleDefinitionTypical example
ProviderDevelops and places an AI system on the market under their own nameAn AI copywriting SaaS company
DeployerUses an AI system under their own authorityA marketing team using that SaaS tool

What is banned under the EU AI Act?

The Act introduces outright prohibitions on several AI practices, effective from February 2025. Marketers should pay particular attention to the first two:

  • Behavioural manipulation — AI systems using subliminal or manipulative techniques that impair a person’s ability to make informed decisions, where this causes significant harm. Aggressive AI-powered persuasion tactics that exploit cognitive biases fall squarely in this category.
  • Exploiting vulnerabilities — targeting specific groups based on age, disability, or socio-economic status to distort their behaviour and cause harm.
  • Social scoring — classifying individuals based on social behaviour or personal traits in ways that lead to detrimental or unjustified treatment.
  • Predictive policing — assessing crime risk based solely on profiling or personality traits.
  • Untargeted biometric scraping — building facial recognition databases by scraping images from the internet or CCTV without targeting specific individuals.
  • Emotion recognition in workplaces and schools — except for medical or safety purposes.
  • Sensitive biometric categorisation — inferring race, political opinions, trade union membership, religious beliefs, or sexual orientation from biometric data.
  • Real-time biometric identification in public spaces — for law enforcement, with only narrow exceptions.

Transparency rules for AI-generated marketing content

This is where the AI Act has the most immediate day-to-day impact on marketing operations. The core obligation: users must know when they are interacting with or viewing AI-generated material.

Chatbots and virtual assistants

If you deploy AI to interact directly with customers, you must inform users they are talking to a machine, unless this is already obvious from the context.

Synthetic images, audio, and video

AI-generated media must be marked in a machine-readable format that makes it detectable as artificially produced. This applies to AI-generated product images, voiceovers, and video ads.

Deep fakes

Content that realistically depicts real people, places, or events must carry a clear and distinguishable disclosure that it has been artificially created or manipulated.

“Deployers who use an AI system to generate or manipulate image, audio or video content that appreciably resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful (deep fakes), should also clearly and distinguishably disclose that the content has been artificially created or manipulated.”

— EU AI Act, official text

AI-written text on matters of public interest

AI-generated text published to inform the public must disclose its AI origin — unless it has undergone substantive human editorial review and a person takes legal and professional responsibility for its publication.

Timing of disclosures

All disclosures must be provided in a clear, distinguishable manner no later than the first interaction or first exposure to the content.

What counts as “human editorial review”?

The disclosure requirement for AI-generated text is waived only when:

  1. The content has undergone substantive human review — not just spellcheck or grammar editing.
  2. A person or entity takes legal and professional responsibility for its publication.

Simple assistive tools that do not substantially change the meaning of content do not qualify.

Key dates and penalties

The Act entered into force in August 2024, with provisions rolling in phases:

  • February 2025 — prohibited AI practices (Article 5) apply.
  • August 2025 — rules for general-purpose AI models apply.
  • August 2026 — most remaining obligations, including deployer transparency rules, apply in full.

Fines for violating prohibited practices can reach €35 million or 7% of global annual turnover, whichever is higher. Other infringements carry fines of up to €15 million or 3% of turnover.

Frequently asked questions

Does the EU AI Act apply to companies based outside the EU?

Yes. If the output of your AI system is intended for use inside the EU, you are within scope as a third-country operator.

Do I need to label every piece of AI-generated content?

For synthetic images, audio, and video: yes, machine-readable labelling is required. For AI-written public-interest text: disclosure is required unless genuine human editorial review has taken place.

Is AI-powered personalisation banned?

Standard personalisation is not prohibited. The ban targets subliminal manipulation that impairs informed decision-making and causes significant harm.

What should I do now to prepare?

Audit which AI tools your team uses and in what contexts. Map whether you are acting as a provider, deployer, or both. Implement disclosure workflows for AI-generated creative and customer-facing AI interactions. Document your human editorial review process for AI-written content.


EU AI Act compliance guide for marketers

Get the full Report


Discover more from The Juicy Marketeer

Subscribe to get the latest posts sent to your email.