25/09/2026
The availability of artificial intelligence has revolutionized the resources available to litigants in person and lawyers alike. Within seconds, generative AI is able to effectively analyse and respond to detailed questions relating to the application of the law to day to day issues. AI models are then able to group the answers into the format of a letter of claim to kickstart a claim.
Following a number of highly publicised instances of AI being used in civil claims, the Civil Justice Council convened a working group in 2025 to review the application of AI in litigation. The group published an interim report in June 2026. From the group’s “direction of travel” Interim Report and Consultation - Use of AI for Preparing Court Documents, it appears likely that a recommendation will be made requiring a declaration within witness statements.
While these recommendations will be no doubt welcomed, it will not interfere with the use of AI at the pre-action stage. In particular the ability of litigants in person to utilise AI to pepper Defendants with letters of claim, complaints, and Data Subject Access Requests (DSARs). We have recently had some success in striking out a claim brought by one such serial Claimant where they had failed to apply for permission to bring a claim based on a previously discontinued claim. However, no such mechanisms are available at the pre-action stage.
When dealing with an AI generated DSAR request, it is useful to remember that the request is only as good as the prompts and questions posed. Where an individual is not sure about their rights, the response generated by them may be based on assumptions that are incorrect, and fabricated law or cases. AI requests are often longer than they need to be, and may request metadata which does not exist, is not accessible, or is not personal data.
That said, if a request is made using AI, it is not automatically invalid. A data controller, be it a public authority, insurer, or small business, will still have to engage with the requestor, and attempt to identify what information the individual is seeking. In extreme cases, it may be possible to say that a request is manifestly excessive or unfounded, but this should be the exception.
Ultimately, the ICO recommends engaging with the individual (and not the bot) to reach agreement on the parameters of a request, and setting expectations early if the request is based on an incorrect premise. If a requestor is adamant that their request is valid, then it may be helpful to use your data protection complaints process to review the matter, and draw a line where agreement cannot be reached. The individual then has the option of making a complaint to the ICO, and you can await further instructions.
Shape your knowledge of public sector delivery — follow our Central & Local Government LinkedIn page.
