Using AI in Your Marketing? From August 2026, Honesty About It Becomes Law in the EU

Most marketers now use AI in some form – drafting copy, generating images, summarising research, and building first-draft reports. What’s changing is the expectation, and in some cases the legal requirement, to be open about it.

From 2 August 2026, the transparency rules of the EU AI Act apply. If you’re a UK business, these rules reach you when your content or chatbot reaches people in the EU. The UK has no equivalent AI law yet, but our regulators are already moving in the same direction. The ASA expects ads not to mislead, however they’re made, and the ban on fake reviews under the Digital Markets, Competition and Consumers Act has been in force since April 2025. The rules are designed to stop AI from being used to deceive people. Here are some tips to keep you on the right side of the law and maintain your customers’ trust.

What the EU rules actually ask for

First, if people interact directly with AI, for example, if your website uses a chatbot to answer customer enquiries, they should be told clearly. Second, certain AI-generated or AI-manipulated content must be labelled. The strictest requirements apply to “deepfakes” (synthetic images, audio or video that look like real people, places or events). AI-written text needs disclosure only when it’s published to inform the public on matters of public interest – and there’s a sensible carve-out for text that’s had genuine human review, where a real person takes editorial responsibility. Raw machine output published unchecked is treated differently from work a human has shaped and stands behind.

One thing you don’t need to worry about: the Act also requires AI providers (OpenAI, Anthropic, Google and the rest) to watermark their outputs in machine-readable form. Your obligations are about how you present content to people.

What this means for normal marketing

You don’t need to write “made by AI” on every social post. Most routine marketing copy isn’t “informing the public on matters of public interest”, and content with genuine human review and editorial ownership sits outside the strictest text rules anyway.

Images and video are where the rules are strictest, and human review isn’t enough. The line to avoid crossing is publishing synthetic content that could pass as real, for example, a fake testimonial video, an AI image presented as a genuine photo of your product, or a customer who doesn’t exist. Those are what the Act is built to catch – and in the UK, the ASA and the fake reviews ban already apply. These were bad practice long before any law mentioned them. One more thing worth knowing: you can’t fix a misleading ad by adding an “AI was used” label. Disclosure of the use of AI doesn’t give you the free rein to publish dishonest content.

Beyond the law: the trust case

This is the part I’d argue matters more for a small business. People are getting better at spotting generic, soulless AI content. Research on consumer attitudes finds that people are more likely to trust brands that use AI openly and well, and switch off from those that are obviously using it without disclosure. Using AI to do the groundwork, while keeping a real human voice and expertise on top, is the ethical approach and the one that works commercially.

Watch for bias

AI learns from historical data, which means it can quietly reproduce or even amplify existing inequalities. A common example: if a model learns that past high-value customers came mostly from one demographic, it may start excluding everyone else from your targeting – narrowing your market and treating people unfairly at the same time. Sense-check who your AI-driven campaigns are reaching and, just as importantly, who they’re leaving out. If the audience looks suspiciously narrow, dig into why.

Take care with the data you feed it

Pasting customer lists, private enquiries or confidential client information into a public AI tool is a data-protection risk and, under UK GDPR, potentially a breach. Treat AI tools like any other third party that touches personal data. Understand where the data goes, use business-grade tools with proper data terms where you can, and keep personal details out of public chatbots unless you’re certain it’s safe. Remember that if a tool is free, then you may be the product.

A simple framework

Three words cover most of it:

Fairness, transparency, accountability.

• Fairness – check outputs for bias and don’t let automation quietly exclude people.
• Transparency – be honest about when and how you use AI, and tell people what happens to their data.
• Accountability – a named human reviews and owns anything that goes out. If you can answer “who checked this, and would we be comfortable explaining how it was made?” for every piece of AI-assisted work, you’re in good shape.

Practical first steps

Write a short, plain internal note on how your business uses AI – what for, what’s off-limits and who reviews what. Keep a record of how AI is used in your content workflow, so that you understand any disclosure obligations. Train whoever uses these tools on the basics of bias and data privacy. And consider a public “how we use AI” note on your website – it costs nothing, answers the question before clients ask it, and signals you’ve got nothing to hide.

The businesses that handle this well won’t be the ones that avoid AI, nor the ones that use it to cut every corner. They’ll be the ones that use it openly – as a tool that makes good people faster, with a human firmly in charge and nothing to hide. That’s a position worth being in well before August, not scrambling for afterwards.

Transparency note: I drafted this article with AI assistance using ChatGPT and Claude, then fact-checked, edited and revised it myself. A human (me) takes full editorial responsibility for it.