Artificial intelligence has rapidly evolved from an emerging technology to an everyday part of agency life.
From brainstorming campaign concepts and generating content to creating visuals and analysing customer behaviour, AI tools are now embedded in the workflows of agencies across the UK. In many cases, clients actively expect agencies to be using AI to improve efficiency, creativity and delivery.
The benefits are undeniable. AI can accelerate production, support ideation, reduce costs and help agencies achieve more with fewer resources.
But as AI becomes increasingly integrated into agency operations, an important question remains: what happens when it gets something wrong?
Consider the following scenarios:
Who is responsible?
The agency, the client or the AI platform itself?
The reality is that there isn’t yet a clear-cut answer. However, for agencies, that uncertainty creates real legal, contractual and insurance considerations.
Most agencies are now using generative AI tools in some capacity, whether internally, within client projects or both.
Platforms such as ChatGPT, Midjourney, Runway and Adobe Firefly are helping agencies streamline processes and improve output. However, alongside these benefits come new areas of exposure that many businesses are only beginning to consider.
Copyright and Intellectual Property Concerns
One of the biggest questions surrounding AI is ownership.
Who owns AI-generated content? If ownership sits with the agency or client, how can they be certain that similar outputs won’t be generated for another user in the future?
AI tools are trained on existing information and datasets. By their very nature, outputs are influenced by pre-existing content. This creates potential challenges where generated material closely resembles existing intellectual property.
As AI adoption increases, copyright disputes and ownership questions are likely to become increasingly common.
Defamation and Reputational Damage
What happens if AI-generated content damages the reputation of a third party?
An inaccurate statement, misleading claim or inappropriate suggestion could cause significant reputational harm if it reaches publication without sufficient review.
The consequences could extend far beyond an unhappy client. Legal action, reputational damage and the loss of valuable client relationships are all potential outcomes.
Data Protection Risks
AI systems learn from the information users provide.
While some platforms offer enhanced privacy controls and enterprise-level protections, no solution is entirely risk-free.
When sensitive, confidential or personal information is entered into AI tools, agencies may expose themselves to data protection concerns, contractual breaches, regulatory scrutiny and potential litigation.
The more valuable the information, the greater the potential consequences if something goes wrong.
Professional Negligence
Many agencies now use AI to support strategic recommendations, content creation and data analysis.
However, if those outputs prove incorrect or lead to client losses, responsibility is unlikely to sit with the software provider.
Clients engage agencies for professional expertise and advice. If AI contributes to a flawed recommendation, agencies may still find themselves accountable.
In many respects, the situation is similar to outsourcing work to a third-party contractor. The client relationship remains with the agency, and so does the responsibility.
Traditionally, agencies have been responsible for the services they provide, rather than the tools they use to deliver them.
AI is beginning to blur that distinction.
Imagine an AI copywriting platform inadvertently reproduces another company’s trademarked slogan. The content is approved, published and later challenged by a competitor.
Or consider a media-buying algorithm that unintentionally discriminates against certain demographics, creating potential regulatory concerns.
Was the problem caused by the software or by the agency using it?
From a legal perspective, clients are unlikely to make that distinction. If AI forms part of your service delivery, responsibility will often still rest with you.
After all, the agency selected the platform, implemented the process and delivered the final outcome.
At present, most Professional Indemnity (PI) insurance policies in the UK continue to cover negligent acts, errors and omissions arising from professional services.
However, relatively few policy wordings currently address AI directly.
This creates uncertainty.
If a claim arises from AI-assisted work, cover may still respond, provided the agency exercised reasonable skill, care and professional judgement.
The challenge lies in determining what constitutes “reasonable” use of AI.
Would insurers consider reliance on unverified AI-generated content reasonable? Or could they argue that greater oversight should have been exercised?
As AI-related claims increase, these questions are likely to become increasingly important.
Exactly how insurers respond remains to be seen, but as AI-related claims emerge and policy wordings evolve, we expect several new conditions, restrictions and exclusions to appear.
Some of the areas we believe insurers may focus on include:
Intellectual Property (IP) Infringement – As AI tools generate content using vast amounts of existing data, there is an increased risk of copyright and intellectual property disputes. Insurers may require agencies to demonstrate that AI-generated content has been reviewed for originality and checked against existing material before cover applies.
Data Protection and Confidentiality – The way agencies use and upload data into AI platforms is likely to come under greater scrutiny. Future policies could include restrictions around the use of client, personal or confidential information in public AI tools, with cover potentially limited where appropriate safeguards have not been followed.
Deliberate or Reckless Use of AI – If an agency knowingly uses unlicensed AI-generated assets or fails to carry out reasonable checks on AI-produced content, insurers may be more inclined to decline claims. While insurers currently assess these situations on a case-by-case basis, we could see clearer policy wording introduced to define what constitutes acceptable AI usage and where cover may not apply.
While regulation and insurance continue to evolve, agencies can take practical steps today to reduce risk and strengthen their position.
Audit Your AI Usage
Identify every process where AI supports or replaces human judgement and determine how those outputs impact clients.
Understand Licence Agreements
Review the terms and conditions of the AI tools you use, paying particular attention to ownership rights, data usage and intellectual property provisions.
Maintain Human Oversight
Ensure AI-generated outputs are reviewed and approved by experienced team members before reaching clients.
This not only improves quality but may also demonstrate reasonable care should a dispute arise.
Protect Client Information
Avoid uploading sensitive, confidential or personal information into public AI platforms unless appropriate safeguards are in place.
Document Your Processes
Create an internal AI policy and maintain records of prompts, edits, reviews and approvals.
A clear audit trail may prove invaluable if a claim is ever made.
Review Client Contracts
Ensure client agreements clearly explain how AI is used, what responsibilities apply and where accountability sits.
Speak to Your Insurance Adviser
Review your Professional Indemnity and Cyber insurance arrangements to understand how AI-assisted work is treated and whether any restrictions apply.
Although insurance markets can be slower to adapt than technology itself, change is already underway.
Insurers increasingly recognise that AI is no longer a niche exposure. It has become a mainstream operational tool used across many professional service sectors.
Over time, we expect to see:
The businesses that prepare early will likely be in a stronger position as both regulation and insurance requirements continue to develop.
AI is transforming the way agencies operate, creating exciting opportunities to improve efficiency, enhance creativity and deliver greater value to clients.
However, innovation also brings new responsibilities.
The legal, contractual and insurance implications of AI are still developing, but clients are unlikely to wait for regulators or insurers to catch up if something goes wrong.
As AI becomes more deeply embedded in agency operations, understanding the risks is just as important as understanding the opportunities. Whether you’re already using AI extensively or are still developing your approach, having the right safeguards in place can make all the difference. If you’d like to explore how your agency can better manage AI-related risks, get in touch with the RiskBox team today. we would be happy to help.
RiskBox are a specialist commercial insurance broker focused on the creative industries, from agencies to tech, media to entertainment. We are truly independent, without any ownership or investment from insurers, therefore our advice is impartial.
You need to load content from reCAPTCHA to submit the form. Please note that doing so will share data with third-party providers.
More Information
