Generative AI is unlikely to change the fundamental principles of complaints handling, investigations or litigation. Organisations will still need to establish the facts, assess evidence objectively and make defensible decisions.
A resident unhappy with a housing decision. A parent challenging a school exclusion. A service user making a complaint. A couple of years ago producing detailed, legally structured submissions often required professional advice, significant research or considerable time and effort. Today, a free generative AI tool can purport to draft a persuasive complaint, appeal or claim in minutes. Generative AI is probabilistic technology which predicts likely responses rather than retrieving verified facts. It can produce information that sounds plausible but is inaccurate.
Whilst the Civil Justice Council is considering whether rule changes are needed to govern the use by lawyers of AI in court documents, its interim update recognises the challenges posed by Litigants in Person (LIPs) using AI. For public service organisations, this matters because Generative AI is lowering the barriers to pursuing complaints, claims, tribunal proceedings and litigation. The result may not be more valid claims, but could mean more claims that appear sophisticated, legally informed and professionally drafted.
Increasingly, LIPs are using generative AI to draft witness statements, prepare case summaries and formulate legal arguments. What was once an emotional, often unstructured complaint can now arrive as a carefully organised submission that mirrors the style and language of a professionally drafted legal document. This creates new risks for organisations tasked with investigating and responding to allegations. The documents may appear more structured and persuasive than traditional complaints, but this does not necessarily mean they are more accurate. The challenge is that these errors can be difficult to identify. A claimant may genuinely believe the output is correct and submit it as part of their complaint or claim.
In some cases, AI generated summaries introduce facts that are not contained within the underlying evidence; one incorrect statement can fundamentally change the perceived strength of a case, creating important questions around evidential integrity. Has a witness account been drafted or substantially rewritten by AI? Does the statement accurately reflect the claimant's recollection, or has AI introduced language, conclusions or assumptions that were not originally present? Are allegations grounded in source documents, or in AI generated interpretations of those documents? Potentially inaccurate material cannot simply be dismissed because it appears to have been generated by AI. Each allegation must still be assessed fairly and investigated appropriately. This increases the time and resource required to manage complaints and claims, particularly where lengthy submissions contain multiple issues that need individually reviewing and addressing.
Responsibility does not rest solely with the claimant using the technology.
Public bodies responding to claims must ensure they are not accepting AI generated material at face value simply because it appears polished or legally sophisticated; all parties have a duty not to mislead the court. In this environment, robust investigation, verification and record keeping become even more critical for public service organisations. The strongest defence against AI generated inaccuracies isn’t more AI, but reliable evidence and a disciplined approach to establishing the facts.
Legal and regulatory landscapes will continue to evolve, but organisations do not need to wait for new rules before taking action.
1. Modernise complaints and claims handling processes.
Organisations should assume that complaints, appeals and claims will become longer, more detailed and legally sophisticated. Complaints handlers and claims teams should focus on the substance of allegations rather than the sophistication of the drafting. A well written document is not necessarily a stronger claim, nor should complex AI generated language obscure the underlying issues that require investigation.
2. Strengthen evidence and investigation practices.
As AI generated content becomes more common, organisations should place renewed emphasis on establishing primary facts and maintaining robust audit trails. Investigators should be alert to the possibility that witness statements, complaint narratives or case summaries may have been drafted or refined using AI tools. This does not make the material unreliable, but it does increase the importance of verifying assertions against contemporaneous records, source documents and objective evidence. Good record keeping, clear chronology management and disciplined evidence review processes will become increasingly valuable safeguards.
3. Invest in AI literacy across risk and governance functions.
Organisations should ensure that those responsible for complaints, investigations, governance and risk understand both the opportunities and limitations of AI. The objective is not to identify whether AI has been used, but to understand how its use may influence the quality, reliability and presentation of information. Teams with AI literacy will be better equipped to assess evidence critically and make proportionate decisions.
The response is not to fear technology. AI has potential to positively improve access to justice. However, greater access does not remove the need for scrutiny, verification and accountability. Organisations most likely to succeed in this evolving environment will be those focussing on strong fundamentals: robust records, disciplined investigations, clear governance and an informed workforce that understands both the benefits and limitations of AI. In a world where convincing narratives are generated in seconds, the ability to evidence what actually happened may become one of an organisation's most important risk controls.
