Generative AI has quickly become part of the marketing toolkit. Businesses use tools such as ChatGPT to help write blog articles, landing pages, product descriptions, emails and social media posts. With the transparency provisions of the EU AI Act applying from 2 August 2026, this raises an obvious question: Do you now need to tell website visitors whenever AI was used to create content? For most businesses using AI as part of a normal content creation process, the answer is no. But there are some important distinctions to understand.
Article 50 of the EU AI Act contains a number of transparency requirements relating to AI-generated and manipulated content.
One of these applies to organisations using AI systems to generate or manipulate text that is published with the purpose of informing the public on matters of public interest. In those circumstances, the organisation may need to disclose that the text has been artificially generated or manipulated.
That sounds broad at first.
However, there is an important exception.
The European Commission explains that this disclosure obligation does not apply where the AI-generated text:
That exception is particularly relevant to marketing teams.
Consider a relatively normal marketing workflow:
The fact that AI contributed to the first draft does not automatically mean the finished webpage needs an "AI-generated" disclaimer.
The company has reviewed the content and has taken responsibility for what it publishes.
This is very different from using an AI system to autonomously create and publish articles without meaningful human editorial involvement.
No.
The AI Act does not introduce a general rule requiring every business to label every sentence, image or webpage that involved AI.
For text, the specific disclosure requirement discussed here concerns AI-generated or manipulated content published for the purpose of informing the public about matters of public interest. The Commission's guidance also confirms the exception for text subject to human review or editorial control.
This means there is an important difference between situations such as:
A marketing team uses AI to help write a new services page. A marketer reviews, edits and approves the text before publication.
An AI disclaimer would generally not be required under this Article 50 provision.
AI helps create a draft article. An editor fact-checks it, changes the content and approves the finished article.
Again, the human-review and editorial-responsibility exception is highly relevant.
A website automatically creates and publishes articles about matters of public interest using generative AI, without meaningful human review before publication.
This is much closer to the situation in which the AI Act's disclosure requirement applies.
Most ordinary commercial website content is also different from the kind of public-interest information specifically addressed by this transparency requirement.
Think about:
There isn't a blanket requirement in Article 50 saying that businesses have to identify all of this as AI-generated simply because generative AI was involved in creating it.
And where a company has a proper editorial process in place, the human-review exception provides another important safeguard.
There is nevertheless a useful lesson for marketing teams in the new rules.
Don't treat AI-generated copy as finished copy.
AI can produce incorrect facts, invent sources, misunderstand context and make claims that your organisation would never consciously approve.
A sensible process is therefore:
AI generates → human reviews → human edits where necessary → company approves → company publishes
The European Commission's guidance refers specifically to a process of human review or editorial control, rather than merely requiring a human to press the publish button.
Businesses using generative AI for content should therefore be able to demonstrate that people remain meaningfully involved in the editorial process.
There is another reason some of the discussion about "labelling AI content" can be confusing.
The AI Act also places obligations on providers of certain generative AI systems to make AI-generated or manipulated content detectable, including through machine-readable marking.
That is not the same thing as requiring every company that uses ChatGPT or another AI tool while writing its website to place a visible disclaimer next to the resulting text.
For marketers and website owners, it is important not to conflate the two requirements.
That becomes more of a question of transparency and brand policy than a blanket legal requirement.
There may be situations where disclosing substantial AI involvement makes sense. A publisher whose content is generated almost entirely automatically, for example, might choose to be particularly transparent about that process.
But adding "This content was created using AI" to every blog article where someone used ChatGPT for brainstorming, research assistance or a first draft is not what the AI Act requires.
It could also create a misleading impression that nobody at the company reviewed or takes responsibility for the content.
A better principle is: Use AI as a tool, maintain meaningful human editorial oversight and take responsibility for everything you publish.
For organisations already using generative AI, the practical response does not need to be complicated.
Establish an internal content policy that makes clear that AI-generated drafts must be reviewed by a person before publication. Make sure factual claims are checked, particularly where the subject is legal, financial, medical, technical or otherwise sensitive.
It should also be clear who ultimately has editorial responsibility for content published on behalf of the organisation.
This is good marketing practice regardless of the AI Act.
The EU AI Act does not mean that every webpage created with the help of ChatGPT now needs an AI disclaimer.
There are transparency requirements for certain AI-generated content, including text published to inform the public about matters of public interest. But there is an important exception where the content has undergone human review or editorial control and a person or organisation assumes editorial responsibility for its publication.
For most marketing teams, the lesson is therefore not "label everything created with AI."
It is:
Make AI part of your editorial process, rather than replacing the editorial process with AI.
Disclaimer: This article provides general information about the EU AI Act from a marketing and website-management perspective and does not constitute legal advice. The application of the AI Act depends on the specific circumstances and use of AI. If you are unsure about your organisation's legal obligations, seek advice from a qualified legal professional.