In short: The advertiser is responsible. The ASA says the CAP Code applies to AI-generated content in the same way as any other content, and that using an AI tool or an automated platform does not move that responsibility away from the advertiser. Check every claim, every likeness and every stereotype before the ad runs, because the ASA will look at the ad, not the tool.
The advertiser is the business whose product is being sold, and the advertiser carries the responsibility. The ASA says the CAP Code applies to AI-generated content in the same way as any other content. An AI tool, an agency or an automated ad platform may make or place the work, but none of them takes the responsibility away from the business that publishes it.
What the ASA says
The ASA set out four points in a June 2026 article on AI and deepfakes:
- The Code is media-neutral. The rules apply however the content was created. Harmful, offensive or socially irresponsible AI images can breach the Code.
- Deepfakes get no exemption from the rules on misleading ads or endorsements. An AI-generated celebrity endorsement that looks real enough to mislead likely breaches the Code.
- The advertiser is responsible for AI output. The ASA says advertisers remain responsible for AI output and should check that it is socially responsible, free from harmful or discriminatory stereotypes, and suitable for the audience and placement.
- Automation does not shift responsibility. In the Stripe & Stare ruling the ad was produced with Google's automated tools and the ASA still held the advertiser primarily responsible for meeting the Code, including where the ad appeared.
What the advertiser has to check
Run these five checks on every AI-made asset before it goes live:
- Accuracy. Check every claim the output makes, including numbers, results, comparisons and before and after images.
- Real people. If the output shows a real person or a celebrity, it must not suggest an endorsement that person never gave. The ASA does not regulate personality rights, so take legal advice before using a public figure's likeness without permission.
- Stereotypes. AI output can carry harmful or discriminatory stereotypes. Look closely at how people, roles and settings are shown.
- Suitability. Check the content against the audience and the place it will run. An ad that suits one channel can be wrong for another.
- Expectation. Ask whether a viewer would be misled about what they are seeing. If they would, change the content or make it clear.
Who signs it off
The ASA's wording is about the advertiser, so the sign-off has to be yours, even if an agency or a freelancer made the work. Put the check in writing. Your contract should say who reviews the output, who approves it and who pays if it has to be pulled. On our own content engine, a person approves every post before it goes out, and the same step belongs in any AI workflow you run.
Common mistakes
- Treating the AI platform's terms as the end of the question. Its terms govern your use of the tool, not whether your ad complies with the Code.
- Publishing a generated testimonial, a figure or a before and after without checking it against reality.
- Leaving the review until the ad is already live. Checks before publication are cheaper than a complaint and a ruling.
- Letting an automated tool place ads with no one reviewing what it generated.
For a fuller view of where AI belongs in your video work, read when to use AI and when to shoot it. If you would rather have a studio produce AI-assisted brand content for you, Granary's AI brand services start at £199 a month +VAT, month to month. The checks above still apply to anything they produce.
Sources
- Advertising Standards Authority, AI and Deepfakes: Four Things Advertisers Need to Know Before They Hit Run (CAP news, 11 June 2026), read 5 November 2026
Related guides
- AI video for small businesses: what's real, what's risky
- When to use AI and when to shoot it for real
- AI in the production workflow, stage by stage
- Voice and likeness: a consent and contract checklist
- How we clone a voice with consent
- All guides and tools
The picture at the top of this page is an illustration, not a photograph of a client or a real job.
Questions we get asked
Who is legally responsible for an AI-generated advert in the UK?
The advertiser is responsible under the CAP Code, according to the ASA. That applies to output from an AI tool and to ads a platform generates or places automatically. The ASA does not regulate personality rights or intellectual property, and it advises legal advice on those points.
Can a business blame the AI tool if an ad breaks the rules?
No. The ASA says automation does not shift responsibility, so the business that publishes the ad still has to meet the CAP Code.
What should I check before running an AI-made ad?
Check the claims for accuracy, any real person or celebrity shown, the people and settings for harmful stereotypes, whether the content suits its audience and placement, and whether a viewer would be misled about what they are seeing.
granary.digital · 01243 884993 · studio@granary.digital · The Old Grainstore, Adsdean, Chichester, West Sussex, PO18 9DN · Prices exclude VAT unless marked inc VAT