We count visits with Google Analytics cookies and Vercel Web Analytics to see which pages help readers most.

Cookie settings
The Assurance Desk

Do UK firms have to label AI in their adverts?

The UK and the EU answer this differently, so where people see your advert changes what you must do.

Not under UK law. There is no blanket legal requirement in the UK to disclose AI in ads, but the usual rules still apply. In the EU, the AI Act has required clear labels on deepfakes since 2 August 2026.

Checked 27 October 2026

Contents
  1. Is there a UK law that says you must label AI in adverts?
  2. When should a UK advert say it used AI?
  3. Does the EU AI Act apply to a UK firm's adverts?
  4. What counts as a deepfake in an advert?
  5. Does AI-written advert copy need a label?
  6. What are the key dates for the EU labelling rules?
  7. What to do
  8. Common questions
  9. Sources

Is there a UK law that says you must label AI in adverts?

No UK law makes you label AI in adverts. In the regulator's words, there is no blanket legal requirement in the UK to disclose the use of AI in ads.

The usual rules still apply. The UK's CAP and BCAP Codes have no AI-specific rules, but the existing rules apply however content is generated, edited or targeted.

There is no separate section of the CAP Code for AI-generated ads. Ads that use deepfakes must meet the rules on misleading advertising, especially on testimonials and endorsements.

A celebrity who appears to endorse a product they have no link with can create problems, including with the CAP Code.

The ASA and CAP published this advice on 29 May 2025. They expect EU and global AI laws to affect the UK too, particularly for brands that advertise globally.

When should a UK advert say it used AI?

For a UK advert, the test is whether people would be misled. The ASA and CAP suggest asking first whether the audience is likely to be misled if the use of AI is not disclosed.

If there is a danger of misleading, the next question is whether the disclosure clarifies the advert's message or contradicts it.

A label does not fix a misleading advert. A disclosure alone is very unlikely to mitigate the harm of a fundamentally misleading message.

An AI image that shows a cosmetic product's effect inaccurately is likely to be materially misleading, even with an AI disclosure.

A disclosure can help in other cases. Saying that a deepfake is only for comedic effect, or that an influencer is AI-generated, could help to negate a misleading impression.

Industry bodies go a little further. In 2023, ISBA and the IPA produced twelve guiding principles on the use of generative AI in advertising.

One principle asks for transparency where AI features prominently in an ad and is unlikely to be obvious to consumers.

The ASA and CAP see issues with requiring every ad made with AI to be labelled. They say they are ready to change their approach as the technology moves.

Should a UK advert say it used AI?

  1. Would people be misled if you did not say AI was used? If not, UK rules need no disclosure, but the advert must still not mislead.
    yes
  2. Would a disclosure clarify the message, not contradict it? If it would contradict the message, a label will not fix it. Change the advert.
    yes
  3. Say that AI was used Make the disclosure clear, in a way that clarifies the message.
The ASA and CAP's two questions, in order. A label never fixes a misleading claim.

Does the EU AI Act apply to a UK firm's adverts?

The EU AI Act can apply to a UK firm's adverts. It covers providers and deployers of AI systems outside the EU, where the output of the AI system is used in the EU.

The deployer is the business that uses the AI system. Where an advertising company uses one, its staff, such as digital animators and content creators, are not separate deployers.

Article 50 of the AI Act applies from 2 August 2026, and from then providers and deployers must comply.

Deployers must disclose deepfakes, and mark text on matters of public interest, with clear and perceivable labels.

Providers, the firms that make the AI tools, must apply a machine-readable mark to AI-generated content, unless the system only assists standard editing.

Labelling AI in adverts: the UK and the EU

UK

  • No blanket legal requirement to disclose AI in ads
  • The CAP and BCAP rules apply however an ad is made
  • Disclose when leaving it out would mislead

EU

  • Article 50 of the AI Act applies from 2 August 2026
  • Deployers label deepfakes with clear and perceivable labels
  • Text on matters of public interest needs a label unless a person reviewed it
Where people see the advert decides which rules apply. A UK firm can be covered by both.

What counts as a deepfake in an advert?

Under the EU AI Act, a deepfake is AI-made or AI-changed image, audio or video. It resembles existing persons, objects, places, entities or events, and would falsely appear authentic or truthful.

Article 50(4) says the deployer shall disclose that such content has been artificially generated or manipulated.

Not every AI edit counts. AI-made background scenes, special effects or technical processing in standard film production are not likely to make content falsely appear authentic.

When a label is needed, people must see it by the time they first see the content. It must be clear and distinguishable.

The hidden machine-readable mark from the AI tool is not enough. Deployers cannot rely on it to meet their disclosure duty.

Evidently artistic, creative, satirical or fictional works still need a disclosure. It must not hamper the display or enjoyment of the work.

Does AI-written advert copy need a label?

AI-written advert copy needs an EU label only when it informs the public on matters of public interest and no person reviewed it.

Matters of public interest include public health, environmental protection, consumer safety and economic or financial developments, among others.

Published text that has undergone human review or editorial control does not need to be labelled.

A quick check is not enough. Spell-checking or grammar correction does not count as human review or editorial control.

What are the key dates for the EU labelling rules?

The key date for the EU labelling rules is 2 August 2026, when Article 50 started to apply.

Deepfakes generated before 2 August 2026 do not have to be labelled retroactively, but labelling them is encouraged.

Tool makers have a grace period for the marking duty until December 2026. It covers generative AI systems placed on the market before 2 August 2026.

The European Commission's guidelines on these duties are dated 20 July 2026. They give definitions, exemptions and practical examples of what is in and out of scope.

The Commission's Code of Practice on AI-generated content is voluntary, but the Article 50 requirements are legal obligations.

The Commission and the AI Board have confirmed that the code is an adequate tool to show compliance. It has a section for deployers on labelling deepfakes and AI-generated text.

By the end of July 2026, about 190 companies and organisations had signed it. Firms that do not sign must show compliance through other equally adequate means.

The EU also created 3 optional icons to label AI-generated content.

Fines can reach 15 million euros or 3% of total worldwide turnover, and proportionality can be taken into account for SMEs.

Labelling AI in adverts: the key dates

  1. 29 May 2025UK The ASA and CAP publish advice on disclosing AI in adverts.
  2. 20 July 2026EU The European Commission's guidelines on the transparency duties.
  3. 2 August 2026EU Article 50 of the AI Act applies.
  4. December 2026EU The grace period for marking ends, for AI systems already on the market.
The UK date is advice from the regulator. The EU dates come from the AI Act and the European Commission.

What to do

  1. Check where people will see the advert, because the EU AI Act covers output used in the EU.
  2. For a UK advert, ask whether people would be misled if you did not say AI was used.
  3. Do not rely on a label to fix a misleading claim.
  4. For an advert seen in the EU, label any deepfake clearly, at the latest when people first see it.
  5. Have a person review AI-written text on matters of public interest, or label it.

Common questions

Does Article 50 apply to my business, and what disclosure text do I need?

It can, if you use AI to make content that people in the EU see. You must disclose that the content was artificially generated or manipulated, in a clear and distinguishable way. The EU created 3 optional icons for this.

Do we have to label adverts made before 2 August 2026?

No. Deepfakes generated before 2 August 2026 do not have to be labelled retroactively, but labelling them is encouraged.

Does a label make a misleading UK advert acceptable?

No. A disclosure alone is very unlikely to mitigate the harm of a fundamentally misleading message.

Does the mark from our AI tool count as the label?

No. Deployers cannot rely on the machine-readable mark from the provider to meet their disclosure duty.

Where next

Sources

  1. asa.org.uk, Disclosure of AI in Advertising: Striking the Balance Between Creativity and Responsibility, ASA, CAP, checked 27 October 2026
  2. European Commission, Transparency obligations under Article 50 of the AI Act, checked 27 October 2026
  3. European Commission, Quick Facts: Transparency rules for AI systems, checked 27 October 2026
  4. EUR-Lex, Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024, checked 27 October 2026
  5. European Commission, Commission publishes guidelines on transparency obligations for providers and deployers of certain AI systems, checked 27 October 2026
  6. asa.org.uk, In at the Deep(fake) End, ASA, CAP, checked 27 October 2026
  7. European Commission, Guidelines on transparency obligations for providers and deployers of certain AI systems, checked 27 October 2026
  8. European Commission, Code of Practice on Transparency of AI-generated Content, checked 27 October 2026

This is an independent editorial guide, not legal advice.