Regulation explainerGrowth, Marketing & Sales

Generative AI in marketing: the disclosure and compliance rules to check before you publish

Generative tools change how marketing assets are made, not the rules for what may be published. Deepfake disclosure under the EU AI Act, endorsement and review rules in the US and UK, copyright limits on AI outputs, likeness rights and platform labelling policies all apply to the finished asset. This page maps those rules and turns them into a review workflow and a checklist.

Reviewed 7 min read

On this page
  1. Why AI-generated marketing needs its own review step
  2. The rules that apply to AI-generated marketing, by jurisdiction
  3. Copyright, ownership and the terms of the tools you use
  4. Likeness, voice and talent rights in synthetic campaigns
  5. Platform labelling policies and provenance metadata
  6. A review workflow for AI-assisted marketing assets
  7. Pre-publication checklist for AI-generated marketing
  8. Questions and answers
  9. Sources

Why AI-generated marketing needs its own review step

Existing review processes check claims, brand fit and legal disclaimers. Generative tools add new questions: does the asset show a person, voice or event that looks real but is not, who owns it, did the tool's license permit commercial use, and does any platform or law require a label? Those questions are easy to miss when an image takes seconds to produce and a campaign needs hundreds of variants.

The safest assumption is that consumer-protection law judges the asset as consumers experience it. A synthetic customer testimonial is a fake testimonial whatever produced it, and an undisclosed paid endorsement by an AI persona is still an undisclosed paid endorsement.

The rules that apply to AI-generated marketing, by jurisdiction

EU AI Act (Regulation (EU) 2024/1689), Article 50

European Union

Applies whenYou publish image, audio or video content that is AI-generated or manipulated and resembles real people, places or events (a deepfake), or AI-generated text published to inform the public on matters of public interest; providers of chatbots and generative systems carry separate duties, and the wider deployer duties are covered in the guide to EU AI Act deployer obligations.1

  • Disclose that deepfake content has been artificially generated or manipulated; for evidently artistic, creative, satirical or fictional work, disclose in a way that does not spoil the work.1
  • Label AI-generated public-interest text unless it has undergone human review or editorial control and someone holds editorial responsibility.1
  • Make disclosures clear and distinguishable at the latest at first exposure. The duties apply from 2 August 2026.2
  • Regulation (EU) 2026/1744 gives providers of generative systems already on the market before that date a four-month transitional period for machine-readable marking only.3

FTC Guides Concerning the Use of Endorsements and Testimonials in Advertising (16 CFR Part 255)

United States

Applies whenAn ad uses an endorsement or testimonial, including from influencers and from endorsers that only appear to be individuals, such as virtual or AI-generated personas.4

  • Endorsements must reflect the honest opinions, findings or experience of the endorser; a persona cannot claim experience nobody had.4
  • Disclose material connections between advertiser and endorser clearly and conspicuously.4
  • Do not imply that an endorser is an independent user when the advertiser created or controls them.

FTC Trade Regulation Rule on the Use of Consumer Reviews and Testimonials (16 CFR Part 465)

United States

Applies whenA business writes, buys, solicits, publishes or suppresses consumer reviews or testimonials; the rule took effect on October 21, 2024.56

  • Do not create, buy or disseminate reviews or testimonials that misrepresent who wrote them or that the reviewer had real experience, which covers AI-written fakes.5
  • Do not offer incentives conditioned on positive or negative sentiment, or publish undisclosed insider reviews.5
  • Do not run company-controlled review sites presented as independent, or suppress negative reviews.5

Digital Markets, Competition and Consumers Act 2024, Part 4 and Schedule 20[^7]

United Kingdom

Applies whenA trader promotes products to UK consumers; the unfair commercial practices regime and its banned practices have applied since 6 April 2025.78

  • Do not submit or commission fake consumer reviews, meaning reviews that appear genuine but are not based on real experience.7
  • Do not publish reviews in a misleading way, and take reasonable and proportionate steps to prevent and remove fake reviews.7
  • Make paid promotion clear in editorial content; hidden advertising is a banned practice.7

US Copyright Office, Copyright and Artificial Intelligence, Part 2: Copyrightability

United States (guidance)

Applies whenYou want to own, register or enforce rights in AI-assisted marketing assets; the report was published on January 29, 2025.9

  • Expect protection only for human authorship: prompts alone are not enough, while human selection, arrangement and modification can be protected.9
  • Keep records of human contributions in case ownership is ever challenged.

Likeness, voice and talent rights in synthetic campaigns

Generating a recognizable person or voice, living or dead, can infringe publicity and personality rights even if no photograph or recording was copied. These rights vary by country and, in the US, by state, so treat any resemblance to a real person as a clearance question rather than a style choice.

When working with real talent, write synthetic use into the contract: which media, for how long, in which markets, whether the likeness can be altered and what payment applies to each reuse. Union agreements and existing talent contracts may restrict digital replicas. The guide to AI voice and likeness protection covers artists' rights in more depth.

Platform labelling policies and provenance metadata

Platforms add their own rules on top of the law. YouTube, for example, requires creators to disclose when they use AI to meaningfully alter or generate realistic content, while unrealistic content and minor edits are exempt.11 Other major platforms publish comparable policies and may label content automatically when they detect provenance data, so check each one before a campaign launches.

Content Credentials based on the C2PA standard attach signed provenance information to a file, recording how it was made and edited.12 They help platforms and audiences verify origin, but metadata can be stripped by screenshots and some upload pipelines, so pair it with a visible label where disclosure is required.

A review workflow for AI-assisted marketing assets

  1. Brief

    Record whether AI tools will be used, for which elements, and whether the output will depict real-seeming people, voices or events.

    Owner
    Campaign owner

  2. Generate with approved tools

    Use tools whose license terms have been reviewed, and log the tool, model version, prompts and source assets for each output.

    Owner
    Creative team

  3. Substantiate claims

    Fact-check every product claim, figure and comparison in generated copy, exactly as you would for human-written copy.

    Owner
    Product marketing

  4. Legal and brand review

    Check endorsements, reviews, likeness, copyright and the tone of voice, and decide which disclosures each market requires.

    Owner
    Legal and brand

  5. Label and attach provenance

    Add visible disclosures where required, apply platform labels and keep Content Credentials intact through export.

    Owner
    Channel owner

  6. Archive the record

    Store prompts, outputs, approvals and license terms with the published asset so you can answer a regulator, platform or rights holder later.

    Owner
    Marketing operations

Pre-publication checklist for AI-generated marketing

0 of 9 checked

Questions and answers

Do we have to label every piece of AI-assisted marketing copy?

Not under the EU AI Act as such. Its text-labelling duty covers AI-generated text published to inform the public on matters of public interest, and it falls away where the text has been through human review and someone holds editorial responsibility. Deepfake images, audio and video are the main deployer obligation. Platform policies, sector rules and your own brand standards may still require labels in other cases.

Can we use an AI-generated influencer in US campaigns?

Yes, but the FTC Endorsement Guides apply to endorsers that merely appear to be people. The persona cannot claim to have used a product when nobody did, the commercial relationship must be disclosed clearly, and you should not imply the persona is an independent consumer. Treat its posts as advertising from your brand, because legally that is what they are.

Is it allowed to use AI to draft replies to customer reviews?

Drafting responses from your business to genuine reviews is generally fine if they are accurate and clearly come from you. What the US review rule and the UK banned practices target is fake reviews themselves: reviews written by or for the business that pose as genuine customer experience, including ones generated by AI, and the suppression or selective publication of real reviews.

Do Content Credentials satisfy disclosure requirements on their own?

Treat them as supporting evidence rather than the disclosure itself. Machine-readable provenance helps platforms detect and label synthetic media, but most viewers never inspect metadata and it can be stripped. Where a law or platform requires that people be told content is synthetic, use a visible or audible label as well.

Sources

  1. Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act) — EUR-Lex · checked 10 October 2026
  2. Quick facts: transparency rules for AI systems — European Commission · checked 10 October 2026
  3. Regulation (EU) 2026/1744 amending Regulation (EU) 2024/1689 (Digital Omnibus on AI) — EUR-Lex · checked 10 October 2026
  4. 16 CFR Part 255: Guides Concerning the Use of Endorsements and Testimonials in Advertising — Electronic Code of Federal Regulations · checked 10 October 2026
  5. 16 CFR Part 465: Use of Consumer Reviews and Testimonials — Electronic Code of Federal Regulations · checked 10 October 2026
  6. Trade Regulation Rule on the Use of Consumer Reviews and Testimonials, final rule (89 FR 68034) — Federal Register via GovInfo · checked 10 October 2026
  7. Digital Markets, Competition and Consumers Act 2024, Schedule 20: banned practices — legislation.gov.uk · checked 10 October 2026
  8. The Digital Markets, Competition and Consumers Act 2024 (Commencement No. 2) Regulations 2025 — legislation.gov.uk · checked 10 October 2026
  9. Copyright and Artificial Intelligence (report series, including Part 2: Copyrightability) — U.S. Copyright Office · checked 10 October 2026
  10. Copyright, Designs and Patents Act 1988, section 9: authorship of work — legislation.gov.uk · checked 10 October 2026
  11. Disclosing use of altered or synthetic content — YouTube Help · checked 10 October 2026
  12. Coalition for Content Provenance and Authenticity (C2PA) — C2PA · checked 10 October 2026

More in Growth, Marketing & Sales

Back to Growth, Marketing & Sales

Next step

Send us one AI-assisted campaign before it goes live

Share the assets, the markets and the tools used. We will map which disclosure, review and rights questions apply and help set up a review workflow your team can repeat.

Review an AI campaign with us