You use AI in your content. But do you need to disclose it? Learn how the EU AI Act is changing the rules on transparency, and what every content team needs to know about AI labelling.
Written by Matej Zachar & Martin Michalik
Your blog draft? ChatGPT. Meta title? Claude. The alt text for your hero image? Yep, AI-generated that too.
In 2026, we’ve moved beyond asking if content teams are using AI. With new EU regulations taking effect, the question every organization now needs to answer is: how are we using AI, and what do we need to disclose?
Let's break it down.
As of 2 August 2026, new EU AI Act transparency and labelling requirements are in effect for certain AI-generated and AI-manipulated content. Alongside these Article 50 requirements, the European Commission has published a voluntary Code of Practice on Transparency of AI-Generated Content, with around 190 organizations signing up ahead of the deadline.
The Code is not legally binding. Instead, it offers a practical blueprint for how organizations can meet the Act’s requirements. It’s a useful guide to what the Commission considers good practice, but it isn't a rulebook.
First, there’s an important distinction to make. Marking and labeling content are not the same thing:
Marking is machine-readable. Think invisible watermarks or signed metadata connected to a piece of content that a system can use to detect that AI was involved. This is aimed primarily at the companies that build AI systems.
Labeling is what a human sees. A visible note on the page saying content was AI-generated or AI-modified. This is aimed primarily at the organizations that publish content.
The Commission has even proposed standardized icons: “AI GENERATED” and “AI MODIFIED”; with guidance on design and placement. Those are optional. You can meet the transparency goal another way.
The rules are narrower than the panic suggests.
The disclosure obligation for text focuses on two things:
New requirements don't require you to stamp an "AI generated" badge on every sentence an LLM has touched.
A meaningful exemption exists.
When content has undergone human editorial review and a named person or organization takes editorial responsibility for it, the text disclosure requirement generally doesn’t apply. For content teams with an established review-and-approve workflow, this may be one of the most important provisions to understand.
The practical takeaway? Human oversight matters. For teams using a headless CMS like Kontent.ai, built-in personalized workflows, permissions, and approval steps can help make that oversight a defined part of the content process, giving teams a clear way to assign responsibility and document editorial review rather than adding another manual compliance task.
That exemption is doing a lot of work, and it’s worth understanding properly rather than labeling everything out of anxiety. Over-labeling has a real cost: if every page carries a disclaimer, the disclaimer stops meaning anything. We’ve collectively been here before with cookie banners.
We’d rather say this plainly than pretend otherwise: nobody has fully settled what the unit of labeling is.
Is it the individual field? The content item? The page? The whole site? A page-level notice and a per-paragraph marking are wildly different amounts of work, and the guidance doesn’t yet resolve that.
The guidelines also don’t detail how to treat content that arrives already touched by AI: translated elsewhere, drafted in another tool, or copied and pasted from a document. Once content is in your system, its history usually isn’t.
Anyone telling you they have this precisely figured out is operating on forecasts, not facts.
The direction is becoming clearer, even though there isn't a single roadmap:
Provenance metadata is consolidating around C2PA Content Credentials, backed by Adobe, Microsoft, Google, OpenAI and others. It’s strongest for images, weaker for video and audio, and largely unsolved for plain text.
Our advice is simple: build the capability now, and be ready to adapt as requirements change.
Start by finding out what you’re actually publishing. Most teams cannot answer the question, “How much of our content involved AI?” That’s the real gap, and it’s a governance problem before it’s ever a legal one.
That last point is the one we’d stress. If the requirements tighten, the teams who already hold this as clean, queryable data will adapt. The teams who don’t will be stuck doing a costly content audit.
Personalized workflows: Keep humans in the loop by designing workflows where content is reviewed and approved by an authorized user before it is published.
Audit log + version history: See what was generated or edited by AI agents, track changes, and maintain a clear record of how content was created and reviewed.
Role-based permissions: Control who can create, edit, approve, and publish content, and prevent AI agents from publishing without explicit authorization.
Certifications: Kontent.ai takes AI governance and security seriously, aligning with leading industry frameworks and certifications, including AI ISO standards, the AI Pact, and the CSA AI Trustworthy Pledge 2025. To learn more about our certifications, visit our trust center.
Rather than introduce a one-size-fits-all AI labeling feature that could become outdated with a single regulatory change, we’re taking a more practical approach. Kontent.ai gives teams the flexibility to meet current requirements, while we continue to update the product as governing bodies provide further guidance. We’ll keep pace with regulatory changes, so your team doesn’t have to.
No. The EU transparency rules focus on deepfakes and on text published to inform the public on matters of public interest. There is also an exemption where content has undergone human editorial review with a named person taking editorial responsibility.
What if we told you there was a way to make your website a place that will always be relevant, no matter the season or the year? Two words—evergreen content. What does evergreen mean in marketing, and how do you make evergreen content? Let’s dive into it.
Lucie Simonova
How can you create a cohesive experience for customers no matter what channel they’re on or what device they’re using? The answer is going omnichannel.
Zaneta Styblova
To structure a blog post, start with a strong headline, write a clear introduction, and break content into short paragraphs. Use descriptive subheadings, add visuals, and format for easy scanning. Don’t forget about linking and filling out the metadata. Want to go into more detail? Dive into this blog.
Lucie Simonova
