FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: Marketing and PR agencies
AI avatars, virtual influencers and voice clones are subject to the same ad rules as human talent, and imitating a real person without consent can lead to legal claims. Agencies carry the risk because they design and produce the synthetic talent.
Part of: AI in Marketing and PR agencies
In short
Virtual influencers, AI avatars and cloned voices are judged under the ASA's media-neutral rules, so marketing must be identifiable and must not mislead about endorsements. IPA/ISBA principles rule out undisclosed deepfakes, ask for openness where AI is prominent and expect respect for personal data and permissions. The UK has no single image right, but people imitated without consent may claim passing off, data misuse or defamation. Media-inclusive PI may respond to some claims, subject to the wording, while data claims and deliberate misuse can fall outside it.
Agencies may build virtual influencers with their own social profiles, create AI avatars to present product videos, and clone voices to produce narration in several languages without rebooking talent. Some campaigns use a synthetic version of a real ambassador; others create a character that looks generic but may resemble someone real. The appeal is cost, speed and control. The risk is that a real person, or a brand, claims the synthetic version trades on them without permission, or that consumers are misled about who is speaking.
The CAP Code has no AI-specific rule, and the ASA applies its media-neutral rules to AI ads. Marketing by an avatar or virtual influencer must be obviously identifiable as advertising in the same way as content by a human influencer, and must not mislead about endorsements or results.
The IPA and ISBA principles are directly relevant. Principle 2 rules out undisclosed deepfakes. Principle 3 asks agencies to be open about AI use where it plays a prominent role and consumers are unlikely to notice, which will often apply to a realistic avatar. Principle 6 requires respect for personal data rights, which matters because a recognisable face or voice can be personal data. Principle 10 expects checks on permissions before content goes live.
The UK has no single standalone image right. Claims by people who say their likeness or voice has been used usually rely on other routes, such as passing off by false endorsement, data protection, defamation or breach of contract. Passing off was also one of the issues tried in Getty Images v Stability AI, though in a different context.
Illustrative scenario: an agency creates a voiceover for a betting client using a voice model trained on a well-known presenter's broadcasts, with no consent. Listeners assume the presenter endorses the brand. The presenter's representatives send a letter before claim alleging false endorsement and misuse of personal data, and the client asks the agency to indemnify it under the services agreement.
Other risks include an avatar that is mistaken for a real employee or customer, a virtual influencer making health claims a human could not lawfully make, and contracts with real talent that do not cover the creation of a digital double, leading to disputes when the clone is reused.
Agency PI that includes media liability may respond to allegations of passing off, defamation or unintentional infringement, subject to its definitions. Cover for data protection claims may sit elsewhere, often in cyber or a specific extension; see GDPR cover within PI. Insurers may treat deliberate use of a person's likeness without consent differently from an honest mistake, and contractual indemnities to clients may go beyond what the policy covers. Compare the options in PI or media liability for AI claims and AI music and voiceover.
Synthetic faces created for still images carry the same lookalike risk, discussed in AI-generated images in ads. If a script makes statements about a real person, our glossary note on defamation cover explains how that part of a policy usually works. More guides are on the marketing and PR agencies hub.
How Apex helps: Apex arranges PI for agencies working with synthetic talent. A broker reviews how your wording treats likeness, passing off and data claims.
No CAP rule specifically requires it. Labelling may be needed to avoid misleading consumers, and IPA/ISBA principles 2 and 3 expect no undisclosed deepfakes and openness where AI plays a prominent role and consumers are unlikely to notice.
Potentially. The UK has no single image right, but a person may claim passing off by false endorsement, misuse of personal data or defamation, depending on the facts. Using a recognisable voice without consent is a real risk.
It may, if the policy includes media or IP liability and the claim falls within its definitions. Data protection elements may need separate cover, and deliberate use without consent may be excluded. Check the wording with your broker.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.
