The short answer
AI-generated advertising is allowed in the UK. The Advertising Standards Authority (ASA) and the Committee of Advertising Practice (CAP) have not written separate rules for it. Instead, the existing UK Advertising Codes apply in full, whatever tools were used to make the ad. If an AI ad misleads, offends or is socially irresponsible, it breaches the Code in exactly the same way a filmed ad would.
In practice that means the question is never “are we allowed to use AI?” but “would this ad mislead or harm anyone if they knew how it was made?”
Do you have to label AI in UK ads?
Not automatically. There is currently no UK rule that requires every AI-assisted ad to carry a label. In guidance published in September 2025, CAP suggested advertisers ask two questions instead:
- Is the audience likely to be misled if the use of AI is not disclosed?
- If so, would a disclosure clarify the ad’s message, or contradict it?
The second question matters more than it looks. CAP is clear that a disclosure is very unlikely to rescue an ad whose core message is misleading. A small “made with AI” note does not fix an image that promises results the product cannot deliver.
Disclosure is most useful where AI people or scenes could otherwise be taken as real: an AI-generated influencer, a synthetic spokesperson, or an obviously comic deepfake that viewers should not mistake for a genuine endorsement.
Where AI campaigns get into trouble
1. Exaggerating what the product does
The classic risk is an AI image or video that shows a product performing better than it really does. CAP’s own example is a cosmetic whose AI-generated “after” shot does not reflect real-world results. The same logic applies to food, fitness, beauty, apps and interfaces: if the frame shows the product, it must show what the product genuinely does.
2. Real people, real likenesses
Deepfakes and look-alikes of real people carry the highest risk. Any AI depiction that could lead viewers to believe a celebrity or public figure genuinely endorses a brand is likely to mislead. The ASA has already ruled against an AI video app whose ad used a well-known singer’s image to invite users to generate kissing videos with celebrities, finding it harmful and offensive. Outside the ASA’s remit, image rights, passing off and intellectual property claims add further exposure.
3. AI “customers” and testimonials
AI UGC can be excellent creative, but an AI character must not be presented as a genuine customer describing a real experience. Since April 2025, the Digital Markets, Competition and Consumers Act has also given the Competition and Markets Authority direct powers to fine businesses for consumer law breaches, including fake reviews. A synthetic “happy customer” is the fastest way to turn a clever campaign into a legal problem.
4. Social responsibility and bias
Generative tools can produce stereotyped or sexualised imagery without being asked to. The ASA has upheld complaints against AI-related ads it found socially irresponsible or offensive, including an AI app ad placed in a mobile game and a fashion ad assembled by automated tools. The advertiser, not the tool or the platform, is responsible for what appears and where it runs.
Enforcement is going automated
The ASA no longer waits only for complaints. Its AI-powered Active Ad Monitoring System is expected to review tens of millions of ads in 2026, and its leadership has said plainly that advertisers using AI will be held responsible for what it produces. Sanctions include published rulings, removal of ads, denial of ad space and referral to Trading Standards. Combined with the CMA’s new fining powers, the cost of getting it wrong has risen sharply.
Running the campaign in Europe too?
If the work will also run in the EU, the EU AI Act adds transparency duties: under Article 50, deepfakes of real people must be disclosed as artificially generated or manipulated, with these obligations scheduled to apply from August 2026. Platforms are moving in the same direction. Meta labels photorealistic AI video, TikTok requires disclosure of realistic AI content, and YouTube asks creators to flag realistic synthetic media. Plan labels and metadata into the edit rather than adding them at the last minute.
A practical checklist for brands
- Treat AI creative exactly like filmed creative: every claim must be true and substantiated.
- Show the real product doing what it really does. Use AI for worlds, not for results.
- Never imply a real person endorses you unless they do, in writing.
- Do not present AI characters as genuine customers or reviewers.
- Ask CAP’s two questions and add a clear disclosure where viewers could be misled.
- Review every output for stereotypes, sexualisation and suitability for where it will run.
- Keep records: prompts, approvals, consents and the evidence behind every claim.
How we approach it at Vane AI
Every Vane AI film starts with the product truth and the brand’s claims, not with the tools. Likenesses are used only with written consent, AI creators are designed as original characters rather than fake customers, and scripts and edits are prepared with the CAP Code in mind before anything goes live. The result is work that can be as cinematic and surprising as the idea demands, and still stand up in front of a regulator.
This article is a general overview of UK advertising regulation as of October 2026 and is not legal advice. For a specific campaign, check the current CAP Code and take advice from a qualified adviser.