The EU AI Act for marketers: what applies when

What the AI Act is

The AI Act is Regulation (EU) 2024/1689, the EU's law on artificial intelligence. Because it is a regulation, it applies directly in every EU country, including Cyprus and Greece. It entered into force on 1 August 2024 and switches on in stages.

The European Commission's AI Act page explains that the law sorts AI by risk:

Unacceptable risk: banned. Examples include social scoring and emotion recognition at work or in schools.
High risk: strict rules before use. Examples include AI that sorts CVs or scores credit.
Transparency risk: people must be told they are dealing with AI, and some AI content must be labelled.
Minimal or no risk: no specific rules, for example spam filters.

Most marketing uses sit in the last two groups. That is good news, but the transparency group is the one that changed this year.

The timeline in plain words

These dates come from the Commission's AI Act Service Desk timeline, which already includes the 2026 changes.

2 February 2025: the bans on unacceptable AI practices apply, plus the general provisions, including AI literacy.
2 August 2025: rules for companies that make general-purpose AI models (the big model makers) apply. Countries had to name their national authorities and set penalties.
2 August 2026: most of the Act applies and enforcement begins. This includes the transparency rules in Article 50.
2 December 2026: a new ban on AI that creates non-consensual sexual deepfakes and child abuse material applies. AI tools that create content and were already on the market before 2 August 2026 must have their machine-readable marking in place by this date.
2 December 2027: high-risk rules apply for uses listed in Annex III, such as employment and education.
2 August 2028: high-risk rules apply for AI built into regulated products, such as machinery.

The December dates come from the "AI Omnibus", a simplification package. The Commission says it entered into force on 27 July 2026. Law firm Goodwin's summary of the changes notes that it is Regulation (EU) 2026/1744, that it delayed the high-risk rules, and that it did not delay the transparency rules. It also softened the AI literacy duty, so check the current text before you plan training around it.

The transparency rules that apply now

Article 50 is the part most marketers need to know. The Commission's Article 50 page sets out four duties:

Chatbots and AI agents. AI systems that talk with people must be built so people know they are dealing with AI, unless it is obvious.
Marking AI content. Providers of tools that generate images, audio, video or text must mark the output in a machine-readable way, so it can be detected as AI-made. This duty falls mainly on the tool makers, such as the companies behind image and text generators.
Emotion recognition and biometric categorisation. Businesses that use these systems must inform the people exposed to them.
Deepfakes and public-interest text. If you publish AI-generated or edited image, audio or video content that looks like real people, places or events, you must disclose that it is artificial. If you publish AI-generated text to inform the public on matters of public interest, you must disclose it too, unless a person reviewed or edited it and someone takes editorial responsibility.

For clearly artistic, creative, satirical or fictional work, the deepfake duty is lighter: you disclose that such content exists in a way that does not spoil the work. The information must be clear and given at the latest at the first interaction or exposure.

On 2 August 2026 the Commission published a news piece, Safer and more transparent AI. It confirms that users must be clearly told when they interact with a chatbot, AI agent or avatar, that the EU has created a set of icons for labelling AI content, and that a voluntary code of practice on AI-generated content can help show compliance. Fines can reach 15 million euro or 3% of global annual turnover, with proportionality for small businesses.

What this means for everyday marketing

Here is how I read it for the work most small businesses do. Treat this as a starting point, not a final answer.

A chatbot on your website or WhatsApp: say clearly at the start that it is an AI assistant, and offer a way to reach a person.
AI product photos and lifestyle images: if an image looks like a real person, a real place or a real event, label it. A realistic AI photo of a "customer" or of a well-known beach in Cyprus is the kind of thing to disclose.
AI voice-overs or videos of people: same rule. If it could pass as real, disclose it.
Blog posts and news-style content: the text duty covers content that informs the public on matters of public interest. If AI drafts it, have a person review and edit it, and stand behind it.
Product descriptions and ad copy written with AI help: these are not named in Article 50's labelling duties. You still need them to be true and accurate under consumer law.

What this means for businesses in Cyprus and Greece

The AI Act applies the same way in Cyprus, Greece and every other EU country. The national authorities enforce it, so local businesses are covered whether they sell locally or across the EU.

In my own work I use AI every day, for drafts, reports and private local tools. The rules do not stop any of that. What they ask is honesty: do not pretend a bot is a person, and do not pass off realistic AI images of real people or places as real. For most small businesses, a clear chatbot greeting and a small "AI-generated" label on realistic visuals cover the main risk.

If you build or sell AI tools yourself, or use AI to screen job applicants or make decisions about customers, you have more to do, and the high-risk dates in 2027 and 2028 matter. Talk to a lawyer.

Questions people ask

Do I need to label every post I write with ChatGPT?

Not every post. The text duty covers AI-generated text published to inform the public on matters of public interest, and it does not apply if a person reviewed or edited it and takes editorial responsibility.

Do I have to add watermarks to AI images myself?

The machine-readable marking duty is on the providers of the AI tools. Your duty, as the business publishing the content, is to disclose deepfakes: realistic content that looks like real people, places or events.

When did the chatbot rule start?

On 2 August 2026, when the Article 50 transparency rules began to apply.

Is this legal advice?

No. This article is a plain-language summary for business owners and is not legal advice. For your own situation, speak to a qualified lawyer.

Want help with this? Book a free 30-minute call.

Sources

Regulation (EU) 2024/1689 (Artificial Intelligence Act), EUR-Lex, June 2024
AI Act, European Commission, Shaping Europe's digital future
Timeline for the Implementation of the EU AI Act, AI Act Service Desk, European Commission
Article 50: Transparency obligations for providers and deployers of certain AI systems, AI Act Service Desk, European Commission
Safer and more transparent AI, European Commission, August 2026
Not Delayed, Not Deferred: EU AI Act Transparency Obligations Are Now in Force, Goodwin, August 2026

The EU AI Act for marketers and small businesses: what applies, and when