--- title: "Labelling AI-generated content – EU AI Act | Sweet Lava" description: "From 2 August 2026 the EU AI Act requires labelling AI-generated content. See which materials you must label and how to prepare your company." url: https://sweetlava.eu/en/blog/labelling-ai-generated-content/ lang: en alternate_pl: https://sweetlava.eu/pl/blog/oznaczanie-tresci-ai/ --- AI # Labelling AI-generated content: what the EU AI Act requires and how to prepare Sweet Lava Team · Published 24 August 2026 · 7 min read **Key answer:** From 2 August 2026, Article 50 of the EU AI Act imposes transparency obligations. A chatbot must tell customers they're talking to AI, deepfakes (realistic AI-generated or AI-altered images, audio and video) need a visible label, and generative tools must mark their outputs in a machine-readable format. Text written with AI that has passed human editorial review needs no label. Breaches carry fines of up to €15 million or 3% of annual worldwide turnover. Table of contents 1. [01The short answer](#the-short-answer) 2. [02Where this comes from: the AI Act in three dates](#where-this-comes-from-the-ai-act-in-three-dates) 3. [03Who's responsible for the label: the tool provider or your company?](#whos-responsible-for-the-label-the-tool-provider-or-your-company) 4. [04Which content needs a label? The four cases of Article 50](#which-content-needs-a-label-the-four-cases-of-article-50) 5. [05Do you have to label AI images in your campaigns?](#do-you-have-to-label-ai-images-in-your-campaigns) 6. [06Platforms have their own rules, independent of the AI Act](#platforms-have-their-own-rules-independent-of-the-ai-act) 7. [07How to label in practice: three layers](#how-to-label-in-practice-three-layers) 8. [08Labelling rules belong in your brand guidelines](#labelling-rules-belong-in-your-brand-guidelines) 9. [09Penalties and supervision](#penalties-and-supervision) 10. [10Where to start: five steps](#where-to-start-five-steps) ## The short answer From 2 August 2026, Article 50 of the EU AI Act applies – the part of the regulation that sets transparency obligations towards the people who see your content. In practice it means four things. The chatbot on your website must say the customer is talking to AI. Realistic images, audio and video generated or altered by AI (so-called deepfakes) need a visible label. Generative tools must mark their outputs in a machine-readable format, which is their providers' duty. Text written with AI that has passed human editorial review needs no label. Below we translate these obligations into everyday marketing situations and show how to prepare your company without paralysing content production. ## Where this comes from: the AI Act in three dates The AI Act is Regulation EU 2024/1689, the world's first regulation covering the AI market as a whole. It entered into force on 1 August 2024, but its provisions were rolled out in stages: from 2 February 2025 the bans on the riskiest practices (such as manipulative social scoring systems), from 2 August 2025 the obligations for providers of general-purpose models, and from 2 August 2026 most of the remaining provisions, including the transparency obligations of Article 50 discussed here. One detail of scope matters a lot: the regulation also covers companies outside the EU whenever their systems or content reach people in the Union. If your campaigns are produced by an agency or a tool from across the ocean, the obligations don't go away. You may recognise this staged roll-out mechanism from another EU regulation we've written about: the [European Accessibility Act](https://sweetlava.eu/en/blog/european-accessibility-act/). ## Who's responsible for the label: the tool provider or your company? The AI Act splits the duties between the provider of an AI system and the deployer – the company that uses the tool. If your team generates graphics in Midjourney, texts in ChatGPT or voiceovers in ElevenLabs, your company is the deployer in the meaning of the regulation. The split looks like this: - **The tool's provider** is responsible for machine-readable marking: the outputs of a generative system have to be marked in a machine-readable format and detectable as artificially generated, for example through metadata or a watermark. - **Your company** is responsible for the layer the audience sees: disclosing a deepfake, telling people they're talking to a chatbot and, in the narrow case described below, disclosing AI-generated text. You can't build the watermark on the provider's behalf, but you do answer for which tools you use and what you publish. From August 2026, choosing tools with solid metadata support is part of legal compliance, not just convenience. ## Which content needs a label? The four cases of Article 50 **Chatbots and voice-based customer service.** AI systems built for direct interaction with people must make it clear the person is dealing with AI, at the latest at first contact. There's an exception where it's obvious to a reasonably well-informed person, but obviousness is a weak thing to rely on: a short note in the chat window settles the matter. We describe how to design these experiences in our article on [AI agents for business](https://sweetlava.eu/en/blog/ai-agents-for-business/). **Deepfakes: realistic images, audio and video.** A deepfake in the meaning of the regulation is content that resembles existing people, objects, places or events and could falsely appear authentic. Such material must carry a clear, distinguishable disclosure that it has been artificially generated or manipulated. For content of an artistic or satirical nature the obligation is lighter: the information can be given in a way that doesn't spoil the experience of the work. **Text on matters of public interest.** The obligation to label AI-generated text applies only to publications that inform the public on matters of public interest, and even then it lapses when the content has passed human review and a specific person or company holds editorial responsibility. A typical company blog or newsletter with human editing is therefore not covered. **Emotion recognition and biometric categorisation.** If a company uses such systems on people, it must inform them. In marketing communication this is rare, but we mention it for completeness because the deadline is the same: 2 August 2026. ## Do you have to label AI images in your campaigns? Three questions settle the answer. Does the material realistically depict people, places, objects or events? Could a viewer mistake it for an authentic photo or recording? Did AI generate or substantially alter the content, rather than merely assist standard editing? Three answers of "yes" mean a deepfake and a visible labelling obligation. On one side of the line: a photorealistic "photo" of a product in the hands of a model who doesn't exist, a generated video of a "customer" praising your service, a voiceover imitating a real person. On the other: a vector illustration, an abstract key visual, an obvious 3D stylisation, plus ordinary retouching, colour correction or background removal, because such operations don't substantially alter the content. Treat borderline cases cautiously: a label costs one line of caption, while its absence can end in lost audience trust and proceedings before a supervisory authority. ## Platforms have their own rules, independent of the AI Act Meta adds "AI info" labels to content it detects or that users declare as AI-generated. YouTube requires creators to disclose realistic synthetic material. TikTok automatically labels content carrying C2PA (Content Credentials) metadata and can remove unlabelled synthetic content under its terms of service. The C2PA standard is provenance metadata attached by tools from Adobe and OpenAI, among others; Google marks the outputs of its models with SynthID. The practical conclusion for you: you need a labelling process anyway, because the law and the platforms you publish on require it in parallel. Better to design it once, properly. ## How to label in practice: three layers **The layer your audience sees.** A short, unambiguous caption next to the material: "Image generated by AI", "Voice generated by AI", or a note in the chat window: "You're talking to an AI assistant". The disclosure should be visible at first contact with the content, not in the terms of service three clicks away. **The technical layer.** Check whether the tools your team uses attach provenance metadata (Content Credentials or an equivalent) and whether your publishing process strips it out, because image compression and exports from graphic tools are the usual culprits. It's the providers' obligation, but verifying it is in your interest. **The process layer.** Decide who on the team rules on labelling, what the exact wording is and where in the material you place it. Write these rules into your brand guidelines, next to the logo and typography rules. Add a clause to contracts with agencies and freelancers: a declaration of AI use and assigned responsibility for the label. While you're at it, take stock of content published earlier: the rules apply from August 2026, but unlabelled realistic AI material from before that date can still breach platform terms and your audience's trust. ## Labelling rules belong in your brand guidelines If your brand already has guidelines for AI tools, which we wrote about in our article on the [AI-ready brand book](https://sweetlava.eu/en/blog/ai-ready-brand-book/), labelling rules are their natural chapter: when we caption material, what the caption says, where it sits in the composition, how the chatbot introduces itself. That way every person and every tool working for the brand applies the same standard, and you don't re-litigate the same doubts with every campaign. We prepare guidelines like these as part of [visual identity projects](https://sweetlava.eu/en/services/branding-design-systems/). ## Penalties and supervision Breaching the transparency obligations carries an administrative fine of up to €15 million or up to 3% of total worldwide annual turnover, whichever is higher. The European Commission is working on guidelines and a code of practice detailing the methods of marking and disclosure, and EU member states are appointing their national supervisory authorities. This article is for information; in ambiguous cases, especially campaigns using people's likenesses, consult a lawyer. ## Where to start: five steps 1. **Take stock of AI content in your company.** List where materials are made with AI: campaign graphics, texts, video, voiceovers, chatbots. Without this list, every next step is guesswork. 2. **Sort materials by obligation.** Realistic images, audio and video: visible label. Text with human editing: no obligation. Illustrations and obvious stylisations: no obligation. 3. **Check your chatbots and voicebots.** Every point where a customer talks to AI needs clear information at first contact. 4. **Set a caption standard and add it to your brand guidelines.** One wording, one placement, one decision process. 5. **Tidy up vendor contracts.** Agencies and freelancers declare AI use, and responsibility for the label is assigned before material goes out. We work with AI daily: we implement it in companies, design with it and prepare brand guidelines that AI tools follow too. If you want to put content labelling in order in your organisation, [tell us about your project](https://sweetlava.eu/en/contact/) and we'll advise where to start. ## Related services - [ AI integration services → ](https://sweetlava.eu/en/services/ai-integration-services/) - [ Brand identity design & design systems → ](https://sweetlava.eu/en/services/branding-design-systems/) ## Related articles AI · 12 June 2026 ## [ Implementing AI in a company, step by step – what our own projects taught us ](https://sweetlava.eu/en/blog/ai-implementation-step-by-step/) How AI implementation works in practice – from a data audit, through a pilot with a measurable goal, to team enablement. With concrete price ranges and a list of situations where AI isn't worth it. AI · 11 June 2026 ## [ AI agents for business – what they are and what they cost ](https://sweetlava.eu/en/blog/ai-agents-for-business/) An AI agent completes multi-step tasks on its own – answering customers, qualifying leads, preparing reports. We explain how it differs from a chatbot, what it costs and where to start. Branding & typography · 5 June 2026 ## [ An AI-ready brand book – how to prepare brand guidelines for AI tools ](https://sweetlava.eu/en/blog/ai-ready-brand-book/) More and more brand materials are created with AI, and a classic PDF brand book is unreadable to the tools. We show how to prepare brand guidelines in a format AI applies directly. Development · 1 June 2026 ## [ GEO/AEO – how do you make your website visible to AI search? ](https://sweetlava.eu/en/blog/geo-aeo-website-visible-to-ai/) More and more buying decisions start with a question to an AI assistant rather than a classic Google search. See what this changes for your business and what to fix first. ## Related terms - [AI hallucinations](https://sweetlava.eu/en/glossary/ai-hallucinations/) - [Chatbot vs AI agent](https://sweetlava.eu/en/glossary/chatbot-vs-ai-agent/) - [E-E-A-T](https://sweetlava.eu/en/glossary/e-e-a-t/) - [LLM (large language model)](https://sweetlava.eu/en/glossary/llm/) ## Frequently asked questions Do you have to label AI-generated content? Some of it, yes. From 2 August 2026 the EU AI Act requires a visible label on deepfakes – realistic images, audio and video generated or altered by AI – and requires telling people when they're talking to a chatbot. Text written with AI that has passed review by a human who takes editorial responsibility for the publication needs no label. Neither do illustrations and graphics nobody would mistake for an authentic photo or recording. When do the AI Act labelling rules apply? From 2 August 2026\. That's when the transparency obligations of Article 50 of the AI Act (Regulation EU 2024/1689) started to apply. The regulation itself entered into force on 1 August 2024, and its provisions apply in stages – the bans on the riskiest practices from 2 February 2025, the obligations for general-purpose model providers from 2 August 2025. Do AI images in advertising need to be labelled? It depends on how they look. A label is required for material that realistically depicts people, places, objects or events and could be mistaken for authentic – for example a photorealistic "photo" of a product in the hands of a model who doesn't exist. An illustration, an abstract graphic or an obvious stylisation isn't a deepfake. Standard retouching and colour correction don't create a labelling obligation either. Do blog posts written with AI need a label? Usually not. The AI Act obligation covers text published to inform the public on matters of public interest, and even then it lapses when the content has passed human review and a natural or legal person holds editorial responsibility for the publication. A company blog with human editing meets that condition. A voluntary label remains a question of your brand's transparency policy. What are the penalties for not labelling AI content? For breaching the transparency obligations of Article 50, the AI Act provides for an administrative fine of up to €15 million or up to 3% of a company's total worldwide annual turnover, whichever is higher. Independently of fines, social platforms can limit the reach of unlabelled synthetic content or remove it under their own terms of service. ## Wondering what AI could change in your company? We implement AI where it makes measurable sense and teach your team to use the tools. We'll also tell you straight when it's not worth it. [ Get in touch ](https://sweetlava.eu/en/contact/?source=blog-ai-content-labeling)