Do you have to label AI-generated ads in the UK?

The short answer

In its May 2025 article on disclosure of AI in advertising, the Advertising Standards Authority (ASA) says there is no blanket legal requirement in the UK to disclose the use of AI in ads. The CAP Code, which the ASA enforces for non-broadcast ads, has no AI-specific rules. What it does have are rules that apply however an ad is made. The ASA describes the Code as media-neutral.

So the question isn’t “did we use AI?” It’s “would anyone be misled?”

The CAP Code rules that matter most

From CAP Code section 3 (misleading advertising):

3.1: ads must not materially mislead, or be likely to.
3.3: ads must not leave out material information, or present it in a way people are unlikely to see.
3.7: you must hold evidence for objective claims before the ad runs.
3.11: ads must not exaggerate what a product can do.
3.44: ads must not contain fake consumer reviews.
3.47: you must hold evidence that any testimonial is genuine, unless it’s obviously fictitious.

When to disclose AI

The ASA suggests two questions:

Would the audience be misled without a disclosure? For example, if an AI-generated person could be taken for a real customer or a real influencer, or AI footage could be mistaken for a real demonstration.
Does the disclosure clarify the ad, or contradict it? A label that explains the ad is fine. A label that tries to cancel out a misleading impression doesn’t work.

The ASA gives the example of an AI-generated image showing a cosmetic effect the product can’t achieve. Adding “made with AI” doesn’t fix it, because the ad still misrepresents the product. On the other hand, a disclosure can help where an ad features an AI-generated presenter, or an obviously comedic deepfake, and people might otherwise assume it’s real.

What a real ruling looks like

In a ruling published on 1 July 2026, the ASA upheld complaints about online ads for a hose attachment. The product images appeared to have been at least partly generated by AI, differed from one another and from the product shown in the video, and one appeared to show electrical components that would have blocked the water flow. Without evidence for the performance claims, the ASA found the ads misleading under rules 3.1, 3.7 and 3.11. The lesson is simple: AI imagery must show the product as it really is.

Deepfakes, celebrities and AI “customers”

Real people. The ASA’s guidance on AI and deepfakes warns that using AI to show a real celebrity or public figure endorsing a product they haven’t endorsed is likely to break the Code. Its article “In at the deep(fake) end” covers the same risks.
Automated tools. The ASA has also made clear that the advertiser stays responsible for an ad even when a platform’s automated tools assembled it.
AI “customers”. An AI-generated person presented as a satisfied customer is a fake testimonial. Since 6 April 2025, fake consumer reviews have been banned outright under the Digital Markets, Competition and Consumers Act 2024, as well as under CAP rule 3.44. The CMA has published fake reviews guidance.

Platform rules are separate

Ad platforms and social networks have their own policies on AI-generated or altered content, and some require labels for realistic synthetic media. These change often, so check the current policy for each platform you advertise on. They apply on top of the CAP Code, not instead of it.

A sensible workflow

The UK advertising trade bodies ISBA and the IPA published twelve guiding principles for the use of generative AI in 2023 (summarised in the ASA’s article). They include being transparent where AI features prominently and isn’t obvious to the audience. In practice:

Start from claims you can prove, and keep the evidence.
Use real product images, or AI images checked carefully against the real product.
Don’t generate real people, and don’t present AI people as real customers.
Where AI people or scenes could be taken for real, add a clear, readable disclosure (for example “AI-generated” on screen).
Check the platform’s own AI labelling settings before you publish.

Quick checklist

Every objective claim is backed by evidence you hold now
The product looks and performs as shown
No real person’s face or voice is used without permission
No AI “customers”, reviews or testimonials
A clear AI disclosure wherever people could think it’s real
The disclosure explains the ad and doesn’t contradict it
Platform AI-label policies checked

FAQ

Is it illegal to use AI in UK ads without a label?

There’s no blanket UK legal requirement to disclose AI in ads, according to the ASA. But the ad must not mislead. If people would be misled without a disclosure, for example because an AI person looks like a real customer, you need one.

Will a “made with AI” label make a misleading ad acceptable?

No. The ASA says a disclosure can’t cure a misleading message, such as AI imagery showing results the product can’t achieve.

Can I use an AI-generated customer testimonial?

No. Presenting an AI-generated person as a real customer is a fake testimonial. Fake consumer reviews are banned under the DMCC Act 2024 and CAP rule 3.44.

Who is responsible if a platform’s AI tool creates the ad?

The advertiser. The ASA has said you remain responsible for your ads even when they are put together by a platform’s automated tools.

Do TikTok, Meta and YouTube have their own AI rules?

Yes, each platform has its own policies on AI-generated or altered content, and they change often. Check the current policy for each one you use, in addition to the CAP Code.

Official sources

Laws and guidance change. Check the sources above for the latest position. For advice on your own circumstances, consult a solicitor.