AI-Generated Ads in the UK: What the ASA Expects

Paper-cut illustration of a shop owner pinning a blank poster to a board while a small robot holds a paintbrush, with a magnifying glass and a blank mask on a stand
Paper-cut illustration of a shop owner pinning a blank poster to a board while a small robot holds a paintbrush, with a magnifying glass and a blank mask on a stand
AI can make the ad; you still answer for it. Illustrative artwork, not an ASA image.

Short answer: There's no blanket UK rule that says you must label every ad made with AI. The Advertising Standards Authority (ASA) says its codes have no AI-specific rules, but the existing rules "apply regardless of how content is generated, edited, or targeted". So an AI image that exaggerates what your product does is misleading in the same way a doctored photo would be, and a small "made with AI" note won't fix a misleading ad. Fake celebrity endorsements made with deepfakes are likely to break the rules. And if an automated ad tool produces something offensive, the ASA says you, the advertiser, are still responsible.

This is a documentation review of the ASA and CAP pages listed under Sources, read on 18 November 2026. It's general information about the advertising codes, not legal advice, and the ASA hasn't reviewed it.

Do you have to say an ad used AI?

Not as a blanket rule. In its guidance on disclosure, the ASA says: "There is no blanket legal requirement in the UK to disclose the use of AI in ads." It also says the CAP and BCAP Codes "do not contain AI-specific rules".

What matters is whether the ad could mislead. The ASA suggests asking two questions:

  1. Is the audience likely to be misled if the use of AI isn't disclosed?
  2. If so, does a disclosure clarify the ad's message, or contradict it?

The ASA points out that advertisers have long used post-production techniques "without a regulatory requirement to always disclose" them, and that AI is often used in a similar way, such as removing unwanted background elements or improving the weather. Showing a result your product can't deliver is different.

A label won't rescue a misleading ad

The CAP Code's first rule on misleading advertising is short: "Marketing communications must not materially mislead or be likely to do so." The ASA says disclosure alone "is very unlikely to mitigate the harm caused by a fundamentally misleading message", and that making a misleading claim and then disclaiming it by saying AI was used would "almost certainly" be against the rules.

For a small business, the risky uses are the ones that show results:

  • Before-and-after images for a beauty, cleaning or repair service.
  • Product photos that make food, rooms or finishes look better than what customers will get.
  • "Customer" photos or testimonials that aren't from real customers.

If an image shows what a customer will get, use a real one or make sure the AI version is accurate.

Deepfakes and fake endorsements

The ASA's 2026 article on AI and deepfakes is blunt: "If an AI-generated celebrity 'endorsement' looks real enough to mislead consumers, it's likely breaching the Code." It adds that personality rights and intellectual property are outside the ASA's remit, so using a public figure's likeness without permission carries legal risks beyond the advertising codes, and you should take legal advice. For who owns the AI images and copy you make, see AI-generated content and UK copyright.

The ASA's disclosure guidance adds that a clear signal can sometimes help, for example making clear that deepfake content is used only for comedic effect, or that an influencer is AI-generated. It says this "will always come down to the specific content and context".

The same tools are used against businesses too. Our guide to AI voice-cloning and deepfake scams covers how to protect your staff.

Automated ad tools don't move the responsibility

Many ad platforms now generate images, headlines and placements for you. The ASA's deepfakes article describes a ruling where an ad was produced using Google's automated tools and still breached the code, and says the ASA found the advertiser "was still primarily responsible for ensuring compliance with the CAP Code". It also warns that "AI bias isn't just a tech problem - it's a compliance problem."

Before an automated campaign runs:

  1. Look at every image and headline the tool has generated, not just the first few.
  2. Check where the ads will appear, especially if your product isn't suitable for children.
  3. Keep proof for any claim the ad makes.
  4. Turn off auto-generated variants you can't review.

A five-point check before you hit "run"

  1. Would a customer be misled if they didn't know AI made this? If yes, change the ad; don't just add a label.
  2. Does it show a result? Use a real or accurate image.
  3. Is anyone recognisable who hasn't agreed? Remove them.
  4. Did a tool generate variants? Review them all.
  5. Can you prove the claims? Keep the evidence on file.

For the wider picture of UK rules that touch AI, see UK AI law in practice. For where AI helps a small firm most, see AI agents for UK small businesses.

Frequently asked questions

Do I have to disclose that an ad was made with AI in the UK?

Not as a blanket rule. The ASA says there is no blanket legal requirement in the UK to disclose the use of AI in ads, and the CAP and BCAP Codes have no AI-specific rules. The usual rules still apply: an ad must not mislead, however it was made. This is general information, not legal advice.

Can a "made with AI" label make an ad acceptable?

Not if the ad is misleading. The ASA says disclosure alone is very unlikely to mitigate the harm caused by a fundamentally misleading message. A label can help in some cases, for example making clear that an influencer is AI-generated, but it depends on the content and the audience.

Can I use an AI-generated celebrity in my ad?

The ASA says that if an AI-generated celebrity endorsement looks real enough to mislead consumers, it's likely breaching the Code. Personality rights and intellectual property are outside the ASA's remit, so take legal advice before using anyone's likeness.

Who is responsible if an automated ad tool creates a problem ad?

The advertiser. In an ASA ruling described in its 2026 deepfakes article, an ad produced using Google's automated tools still breached the Code, and the ASA found the advertiser was still primarily responsible for compliance.

Sources

  • ASA, "Disclosure of AI in Advertising: Striking the Balance Between Creativity and Responsibility", opened 18 November 2026: https://www.asa.org.uk/news/disclosure-of-ai-in-advertising-striking-the-balance-between-creativity-and-responsibility.html
  • ASA, "AI and Deepfakes: Four Things Advertisers Need to Know Before They Hit 'Run'", 11 June 2026, opened 18 November 2026: https://www.asa.org.uk/news/ai-and-deepfakes-four-things-advertisers-need-to-know-before-they-hit-run.html
  • CAP Code, Section 3: Misleading advertising (rule 3.1), opened 18 November 2026: https://www.asa.org.uk/type/non_broadcast/code_section/03.html

Last checked: 18 November 2026 (Europe/London). Evidence: documentation review of ASA and CAP pages. No ads were tested. General information, not legal advice.

Scroll to Top