AI insurance

AI insurance for marketing and creative agencies

Reviewed by the Apex broking team · Last reviewed 2026-10-26

AI insurance for agencies covers the claims a marketing, creative or content agency can face when it uses generative AI in client work, such as copyright, defamation and misleading ads. It matters because clients now ask agencies to promise that AI-assisted work is original and lawful, and that promise sits with the agency, not the AI tool.

Part of: AI insurance UK

In short

Agencies using generative AI face claims over copyright and trade marks in AI output, misleading or defamatory copy, deepfakes, fake reviews and misuse of client data. UK law gives computer-generated works a limited form of copyright under section 9(3) CDPA, but the government proposed removing it in March 2026, so ownership warranties need care. The ASA says advertisers stay responsible for AI output, and CAP Code rule 1.8 binds agencies too. The CMA can fine for fake reviews. PI, IP infringement and cyber cover usually do the heavy lifting, subject to the policy terms.

Where agency AI claims come from

Agencies now use generative AI for concepts, copy, images, video, voice, social content and ad targeting. Each use carries a different claim. The examples are illustrative.

AI use	What can go wrong (illustrative)	Who might claim or act
Image and video generation	An output closely resembles a photographer's work or another brand's mark	The rights holder, then your client
AI copywriting	An untrue product claim, or a false statement about a competitor	The ASA, the CMA, the competitor, your client
Synthetic voices, avatars and deepfakes	A real person's likeness implies an endorsement they never gave	The person, the ASA
AI-written reviews or testimonials	Reviews that do not reflect genuine customer experience	The CMA, your client
Client data in AI tools	Confidential or personal data used in breach of contract or data protection law	Your client, its customers, the ICO

Who owns AI-generated creative: CDPA section 9(3) and the March 2026 report

Clients usually expect you to assign copyright in deliverables. With AI-generated work there may be less to assign than they assume.

Section 9(3) of the Copyright, Designs and Patents Act 1988 says the author of a computer-generated work is "the person by whom the arrangements necessary for the creation of the work are undertaken". Section 178 defines computer-generated as having no human author, and under section 12(7) such copyright lasts 50 years from the end of the year the work was made.
Whether a wholly AI-generated image or text is original enough to attract copyright at all is uncertain in UK law.
The government's Report on Copyright and Artificial Intelligence, presented to Parliament in March 2026 under section 136 of the Data (Use and Access) Act 2025, proposes that protection for wholly computer-generated works "should be removed" in the absence of evidence of its value, while works created with AI assistance stay protected. It also proposes working with industry on labelling AI-generated content and exploring a new right over digital replicas (HSF Kramer summary). These are proposals at October 2026, not law.

In practice, a warranty that "the agency owns and assigns all rights" is harder to stand behind for heavily AI-generated work. Meaningful human creative input, and records of prompts, edits and selection, help show authorship. For infringement risk, including Getty Images v Stability AI, see AI IP infringement insurance.

Client contract terms agencies are being asked to sign

Client agreements and procurement terms increasingly include:

disclosure of AI use, or a ban on AI without written approval;
warranties that deliverables are original and do not infringe third-party rights;
IP indemnities, sometimes uncapped;
bans on putting client confidential information into public AI tools or letting it be used for training;
compliance with the CAP Code and consumer law.

PI and IP covers commonly exclude liability you take on by contract beyond what you would owe anyway, so an uncapped indemnity can be partly uninsured. Negotiate caps and knowledge qualifiers, and tell your insurer about unusual terms. See AI contract clauses and indemnities.

ASA and CAP rules on AI in ads

There are no AI-specific advertising rules in the UK. The ASA's guidance of 11 June 2026 says the CAP Code is media-neutral, that "If an AI tool produces something problematic, the advertiser remains responsible for any output used in their ads", and that an AI-generated celebrity "endorsement" realistic enough to mislead is likely to breach the Code. In the Stripe & Stare ruling, an ad assembled with automated tools still left the advertiser primarily responsible.

CAP has also said that disclosing AI use is unlikely to cure an ad that is misleading, giving the example of an AI image showing a cosmetic effect that does not reflect real-world results (RPC summary).

Agencies are not off the hook. CAP Code rule 1.8 puts primary responsibility on marketers but says agencies "also accept an obligation to abide by the Code". An upheld ruling can leave a client with the cost of pulling a campaign, and some clients then look to the agency to recover it.

Misleading claims and fake reviews under the DMCC Act 2024

Since 6 April 2025 the CMA can enforce consumer law directly and fine businesses up to 10% of global turnover (CMS). Section 225 of the Digital Markets, Competition and Consumers Act 2024 prohibits unfair commercial practices, and its definition of trader includes "a person acting in the name of, or on behalf of" a trader, which can reach agencies.

Paragraph 13 of Schedule 20 bans submitting or commissioning fake consumer reviews, and offering services to traders to do so. A fake consumer review is one that purports to be, but is not, based on a person's genuine experience. AI-written testimonials fall squarely into this. Fines are often not insurable: in Safeway Stores v Twigger [2010] EWCA Civ 1472 the Court of Appeal held that a company could not pass its own competition penalty on to its former staff. See are AI regulatory fines insurable.

Defamation, likeness and data

Defamation. Generative AI can produce false statements about real people and companies. Under section 1 of the Defamation Act 2013 a statement is not defamatory unless it has caused or is likely to cause serious harm, which for a business means serious financial loss. Many agency PI policies include defamation cover; check yours. See does PI cover defamation.
Likeness. The ASA notes it does not regulate personality rights or IP and advises legal advice before using a public figure's likeness.
Client data. If you process client customer data, Article 28 of the UK GDPR requires a processor contract, and client agreements usually add confidentiality duties. Pasting that data into an AI tool can breach both.
EU audiences. EU AI Act transparency duties, including disclosure of deepfakes by deployers, have applied since 2 August 2026 (Kingsley Napley). See the EU AI Act for UK businesses.

Which insurance responds

Claim (illustrative)	Cover that may respond, subject to the policy terms
Client sues because an AI-assisted campaign was wrong or late and cost it money	Professional indemnity
Photographer or brand alleges an AI image copies its work or mark	IP infringement, or PI with an IP extension
A competitor sues over an AI-written comparative claim	PI or media liability with defamation cover
Client data leaks through an AI tool	Cyber for response costs; PI for the client's claim
CMA fine for fake reviews	Generally not insurable; defence costs may be covered under some wordings

Check for AI exclusions and whether your business description mentions AI work; see affirmative AI cover vs silent AI, cyber and PI overlap for AI content, PI vs media liability for marketing and PI for marketing agencies.

How Apex's AI insurance scheme works for agencies

Apex has an AI insurance scheme with an insurer in which your specific AI work, for example "creative and content agency producing AI-generated imagery, copy and synthetic voice for retail clients" (illustrative), is written into the policy as insured activity rather than left to a generic business description. Cover options are professional indemnity, cyber, product / public liability and IP infringement, so the main agency exposures can sit in one policy, and cover for US exposure is available, subject to the insurer's underwriting and the policy terms. To quote, we need a precise description of your AI work, your address, your turnover (including how much comes from the USA and Canada) and the limit of indemnity you want; you can start an AI insurance quote online in a few minutes. See AI insurance UK.

Controls that help with clients and insurers

A written AI policy: approved tools, enterprise terms and what data may go into them.
Human review of every AI output before it reaches a client or the public, with claims checked against evidence.
Similarity and clearance checks on AI imagery, and consent for any real person's likeness or voice.
Records of prompts, edits and approvals for each deliverable.
A contract register showing which clients have AI clauses, caps and indemnities, and which are in the USA or Canada (see US cover for UK AI firms).

Sources

Frequently asked

Does my agency's PI insurance cover work created with AI?

It may, if the work falls within the business described in your policy and no AI exclusion applies. Many policies were written before generative AI and say nothing about it. Check the business description, IP and defamation cover and exclusions, and tell your insurer how you use AI. Cover is always subject to the policy terms.

Who owns copyright in AI-generated designs in the UK?

For a computer-generated work with no human author, section 9(3) CDPA gives authorship to the person who made the arrangements for its creation, which could be the agency or the client. Whether such work is original enough is uncertain, and the government proposed in March 2026 to remove this protection. Human creative input strengthens a claim.

Is the agency or the client liable for a misleading AI-generated ad?

Both can be. The ASA treats the advertiser as primarily responsible, but CAP Code rule 1.8 says agencies also accept an obligation to abide by the Code. Under the DMCC Act 2024 a trader includes someone acting on a trader's behalf. Clients may also claim against agencies under their contract.

Do we have to label AI-generated ads in the UK?

There is no general UK rule requiring it. CAP's position is that you must not mislead, and disclosing AI use will not cure an ad that is misleading. For EU audiences, EU AI Act transparency duties on deepfakes have applied since 2 August 2026. Check client contracts, which may require disclosure.

Can an agency be fined for AI-generated fake reviews?

It can be at risk. Schedule 20 of the DMCC Act 2024 bans submitting or commissioning fake consumer reviews and offering services to traders to do so, and since 6 April 2025 the CMA can fine businesses directly. Fines are generally not insurable, though some policies cover defence costs.

Get a quote for AI insurance

Tell us what your AI work is, where you are based, your turnover and the limit you need. It takes a few minutes online, and a broker will come back to you with options. Or call 0117 325 0027.

Start an AI insurance quote Request a callback

Apex 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, not advice on your individual circumstances, and it does not guarantee that cover will be available or on what terms.