Which Authority Oversees the AI Act in Greece?

The honest answer is that it is still being settled. Under the EU AI Act, Regulation (EU) 2024/1689, each Member State must designate its own national competent authorities and market-surveillance authorities to supervise and enforce the Act. In Greece, as in a number of other countries, those bodies are still being designated during 2025 and 2026. Why there is no single fixed name yet The AI Act is directly applicable across the Union, but it leaves enforcement to national authorities. Each country decides which existing regulator, or new body, takes on the role, and it notifies that choice at national level. Because this designation work is ongoing, the precise Greek authority may be confirmed or adjusted as the framework beds in. What this means for your shop Your duties do not wait for the authority to be named — the transparency rules apply directly from the Regulation. You should still disclose AI chatbots and label AI-generated content, in Greek for the Greek market. Keep an eye on which national authority Greece designates, so you know who to look to for supervision and guidance. For the broader phased picture, see our complete AI Act guide. Since the national authorities are still being designated, treat the enforcement detail as developing. This answer is educational, not legal advice.

Must Romanian Shops Disclose AI Chatbots?

In short, yes. If you run an AI chatbot on a shop selling into Romania, you should tell customers they are dealing with a machine rather than a person, unless that is already obvious from the context. The duty comes from the EU AI Act, Regulation (EU) 2024/1689, which applies directly in Romania as an EU rule. Where the duty comes from Article 50 of the AI Act sets out transparency obligations. One of them is that people interacting with an AI system should know they are doing so. For a shop, that means a chatbot or virtual assistant should make clear it is automated. Most shops are "deployers" of AI — they use it rather than build it — and this disclosure is a deployer-level, practical step, not a heavy technical burden. What "disclose" looks like in practice A short line near the chat window saying customers are speaking to an automated assistant. Wording that is clear and easy to see, not buried in the small print. The disclosure available in Romanian for your Romanian customers. The Romanian angle The rule is the same in Romania as across the Union, because it comes from an EU Regulation. Romania adds the local layer: national authorities to supervise the Act — still being designated during 2025 and 2026 — and the expectation that consumer information reaches Romanian buyers in Romanian. This answer is educational, not legal advice.

Which Swedish Authority Oversees the AI Act?

The EU AI Act (Regulation (EU) 2024/1689) is enforced by national authorities that each Member State designates. In Sweden these bodies are still being designated, so at the time of writing there is no single, settled answer to exactly which authority you would deal with. This is normal during the Act's phased roll-out and does not change your obligations as a shop. How oversight is structured The Act is uniform across the EU, but enforcement is left to the Member States. Each country designates national competent authorities and a market-surveillance authority. In Sweden this designation process is under way, and existing supervisors are expected to take on roles once the arrangements are confirmed. Because it is not yet finalised, the precise allocation may still change. What this means for a shop You do not need to wait for a named authority to start complying — Article 50 transparency already applies.Disclose AI chatbots and label AI-generated media now, in Swedish.Keep a short internal record of how you comply, so you can respond easily once the supervisor is confirmed. In short, the "which authority" question is still being answered, but the duty to be transparent about your AI is fixed and applies today. This article is educational, not legal advice.

Do Polish Shops Label AI-Generated Images?

Yes. If your Polish webshop uses images that were generated or meaningfully altered by artificial intelligence, the EU AI Act (Regulation (EU) 2024/1689) expects you to label them. The relevant rule is the transparency obligation in Article 50, which applies directly in Poland because the AI Act is an EU regulation. What the rule covers Article 50 asks that AI-generated or manipulated image, audio and video content is marked as artificial, so people are not misled into thinking synthetic media is a genuine photograph. For a shop, this typically reaches product visuals, lifestyle imagery and marketing graphics that an AI tool has created. Realistic AI depictions of real people or places — deep fakes — must also be disclosed as artificial. How to comply in practice Add a visible note where AI-generated images appear, in Polish so customers understand it.Be especially careful with images that could be mistaken for real photographs of the actual product.Keep a record of which visuals are AI-generated and how they are labelled. A short, honest label is enough. The goal is that a customer can tell at a glance that the image was made or altered by AI. Who enforces it? Enforcement rests with national authorities that Poland designates. That designation is still being finalised, so the practical advice is to label AI-generated images now rather than wait for a named regulator. This article is educational, not legal advice.

Who Supervises the AI Act in Belgium?

The EU AI Act (Regulation (EU) 2024/1689) is enforced by national authorities that each Member State designates. In Belgium these bodies are still being set up, so at the time of writing there is no single, settled answer to exactly which office you would deal with. This is normal across the EU during the Act's roll-out and does not change your underlying obligations. How supervision is organised The Act itself is uniform EU-wide, but it leaves enforcement to the Member States. Each country designates national competent authorities and market-surveillance authorities. In Belgium's federal system this involves coordination between federal and, potentially, regional bodies, and existing supervisors are expected to take on parts of the role. Because the designations are still being finalised, the precise allocation of responsibilities may shift. What this means for a shop You do not need to wait for a named authority before complying — the transparency duties in Article 50 already apply.Disclose AI chatbots and label AI-generated media now, in the appropriate regional language.Keep a short internal record of how you comply, so you can respond easily once the supervisory picture settles. In short, the "who" is still being decided, but the "what" — being transparent about your AI — is fixed and applies today. This article is educational, not legal advice.

Guide

Labelling AI-Generated Product Photos and Descriptions

By Compliance Editorial Team Updated July 5, 2026 4 min read

AI Act

Artificial intelligence has quietly become part of everyday online retail. Shops generate product imagery, write descriptions, and answer customer questions with AI tools. The EU AI Act, Regulation (EU) 2024/1689, brings transparency duties that touch all three of these uses. This guide explains what Article 50 asks of an online shop and how to label AI-generated content so your customers are not misled.

Are you a deployer under the AI Act?

Most shops do not build their own AI models. They use third-party tools to generate an image, draft a description or power a chatbot. In the language of the AI Act, using an AI system in the course of your business makes you a deployer, and deployers carry the Article 50 transparency duties for the content and interactions they put in front of customers. In other words, buying the tool does not outsource the duty to disclose. You can read our summary on the AI Act law page and the fuller complete AI Act guide.

What Article 50 requires

Article 50 sets out several transparency duties that map neatly onto shop scenarios:

Tell users when they are interacting with an AI system rather than a person, for example a support chatbot.
Mark AI-generated or manipulated image, audio and video content as artificially generated.
Disclose AI-generated or manipulated text where it informs the public on matters of public interest, and disclose deep fakes.

The disclosures should be clear and given at the right moment: when the customer starts chatting, or when they view the content in question. Burying a single line in your terms and conditions is not the spirit of the rule.

AI-generated product photos

Synthetic imagery is where shops most often stray. If you use AI to create a lifestyle scene, place a product in a setting it was never photographed in, or generate a model wearing your item, that image is artificially generated and falls within the marking duty for synthetic visual content. Label it plainly, for example with a visible caption such as “AI-generated image”, so a shopper is not led to believe they are seeing a genuine photograph of the actual product.

Beyond the AI Act, misleading imagery can also raise separate consumer-protection concerns, so accurate labelling protects you on more than one front. If the picture materially changes what the customer thinks they are buying, transparency is not optional politeness, it is the point.

AI-written product descriptions

Text is more nuanced. The Article 50 text-marking duty is framed around AI-generated text that informs the public on matters of public interest, which a straightforward product description usually is not. That means the strict marking obligation frequently will not attach to ordinary catalogue copy. Even so, the safer and more customer-friendly approach is to keep AI-drafted descriptions accurate, review them before publishing, and avoid presenting machine-written claims as tested human judgement.

Where AI copy strays into areas the public relies on as information, or where you are unsure, a brief disclosure costs little and removes the risk. We cover the chatbot side of this in our FAQ on AI transparency for online stores.

Chatbots and customer interactions

If a shopper lands in a chat window staffed by AI, tell them so. A short line such as “You are chatting with an automated assistant” at the start of the conversation satisfies the interaction-disclosure duty and sets honest expectations. If a human takes over later, it is good practice to say that too.

A practical labelling checklist

Add a visible “AI-generated image” caption to any synthetic product photo.
Announce chatbot interactions up front, in plain language.
Review AI-drafted descriptions for accuracy before they go live.
Keep a note of which listings use AI content, so you can update labels consistently.

Transparency is cheap and trust is expensive. Telling customers what is AI-made costs a caption and buys you credibility.

This article is educational and does not constitute legal advice.

Are you a deployer under the AI Act?

What Article 50 requires

AI-generated product photos

AI-written product descriptions

Chatbots and customer interactions

A practical labelling checklist

Related reading

The EU AI Act in the Czech Republic: What Online Shops Must Disclose

The EU AI Act in Denmark: What Online Shops Must Disclose

The EU AI Act in Hungary: What Online Shops Must Disclose

The EU AI Act in Ireland: What Online Shops Must Disclose