The Civil Justice Council (CJC) has issued its Final Report on the Use of AI for Preparing Court Documents, setting out 15 recommendations for the safe and responsible use of artificial intelligence (AI) in civil court proceedings. We examine the key recommendations and what they could mean for legal practitioners, the preparation of court documents and future civil litigation practice.

Key Recommendations

  • No prohibition on the use of AI tools, but a focus on their effects
  • Witnesses of fact to certify that AI did not generate or influence the substantive content of their statement
  • Experts to declare any use which did generate or influence the substantive content of their report, and give information to allow scrutiny
  • Disclosure – no changes proposed, but subject to Disclosure Working Group's ongoing review
  • Written Advocacy – no changes, but emphasis on lawyers' personal responsibility
  • Guidance recommended for litigants in person, regulators encouraged to cooperate, and a proposed new AI Practice Direction

Why did the CJC review the use of AI in court documents?

The CJC is a body whose primary role is to provide recommendations on how civil justice can be made more accessible, fair and efficient. In January 2025, it established a working group to examine the use of AI for preparing court documents, and to consider whether rules are needed to govern AI use in civil proceedings. The Terms of Reference were restricted to the use of AI by legal representatives – use of AI by the Judiciary and unrepresented litigants were out of scope.

The working group published an Interim Report and Consultation in February 2026, and an update on its consultation findings and emerging direction of travel in June 2026. The Final Report, issued on 29 October 2026, sets out the CJC's conclusions and recommendations for further action in a number of key areas.

New guidance on the responsible use of AI in civil proceedings

The Report recommends that legal regulators should remain primarily responsible for providing guidance to the legal profession on the responsible use of AI. While regulators including the Solicitors Regulation Authority and Bar Standards Board have already issued guidance on the topic, the Report also counsels regulators to work together collaboratively to ensure that such guidance is mutually consistent, and to explore the possibility of issuing joint guidance.

Alongside regulatory guidance though, the CJC recommends that the Civil Procedure Rule Committee issue a new Practice Direction, written in plain language, to codify and consolidate the body of case law which has built up around the use of AI, and sanctions for its misuse.

AI use by litigants in person

Despite being outside the CJC's original Terms of Reference, the consultation phase appears to have drawn a considerable level of concern among consultees about the use of AI by unrepresented litigants. While use of AI by litigants in person may facilitate access to justice, the Report notes that harm can nonetheless be caused to litigants on both sides, and the court through injudicious use (including where its use leads to voluminous court documents).

It is striking therefore that, while out of scope, the Working Group considered that the CJC should issue clear, accessible guidance for litigants in person as a matter of urgency – and that guidance should be reflected in any new Practice Direction on AI use.

AI in written advocacy– existing professional duties are sufficient

Considering written advocacy documents prepared by legal representatives (e.g. statements of case and skeleton arguments), the Working Group concludes that existing professional duties owed by legal representatives to the court and regulators are sufficient to police the use of AI.

Accordingly, it does not recommend that legal representatives should have to declare any use of AI in preparing such documents, though they remain personally responsible for their content.

AI in disclosure – no change recommended

The Report notes that the use of technology is now well-established in the disclosure process for some time and concludes that no reforms are required concerning any use of AI in disclosure. However, that is subject to any recommendations from the Disclosure Review Working Group (expected later in 2026 or early 2027), which is currently working separately on a review of Practice Direction 57AD, including on the topic of whether recent technological developments are impacting the disclosure process.

AI in witness statements – a new declaration proposed

When it comes to witness statements however, the Working Group recommends requiring the introduction of a specific declaration that any AI used in its preparation has not generated or influenced the substantive content of the statement. 'Generated' is intended to cover drafting which is suggested by an AI (rather than capturing the witness's own words), while 'influenced' captures a process by which although AI does not produce the content, it affects its selection, analysis, development or evaluation.

This proposed declaration would apply to witness statements of fact, but not to witness statements given by legal representatives on procedural matters, where existing professional safeguards and statements of truth are considered sufficient.

AI Translation of court documents: human certification remains key

AI also has a potential role to play in the translation of court documents, including witness statements. While the Working Group does not believe any new rules are required to govern this use of AI, it notes it is already the case that any translation of a witness statement must be accompanied by the document in its original language, and a certification of accuracy from a human translator. It recommends however that Civil Procedure Rules (CPR) 32 and its Practice Direction be updated to make it express that this requirement applies equally to AI translations.

AI in expert report – Disclosure where AI influences content

By contrast with witness statements, where the Working Group recommends a negative declaration, in the case of expert reports they recommend a positive declaration be given by exception, where the expert has used AI to generate or influence the content of their report. Where this declaration is given, the expert should disclose details of the AI tool used and the methodology, to allow the court and opposing parties to scrutinise the expert evidence effectively.

In tandem with this, the Working Group notes that the CJC is separately reviewing its 2014 Guidance for the Instruction of Experts in Civil Claims, and recommends that this working party should introduce best practice guidance on AI as part of its review.

A risk-based approach to AI in civil litigation

The philosophy behind these recommendations is pragmatic and outcomes-focused – the CJC has not sought to prohibit the use of AI tools, generally or specifically, but has focused instead on identifying their potential harms to the integrity of the litigation process, and how they can be mitigated proportionately.

That has led to minimal intervention in lower-risk categories like written advocacy, where professional duties (and sanctions) are considered effective, and a more prescriptive declaration proposal for witness statements of fact, to emphasise the importance of the witness's own words for the integrity of the process.

The Working Group's approach also avoids overly technical definitions and categorisations of AI technologies. In part, that is because the rapid pace of development means that overly specific definitions may quickly be rendered obsolete, and because the increasing integration of AI tools into a range of products makes its use harder to ascertain. But it is also informed by consultation responses that caution a focus on the tools and their categorisation (rather than their results and their impact on the litigation) could provide a new battleground for unnecessary and costly satellite litigation.

While the recommendations will now need to be considered further by the Civil Procedure Rule Committee, we can perhaps expect to see changes in time for the April CPR Update.

For more information on the use of AI in litigation, please contact Sean Adams, Alex Brodie, Michael Carter or Christopher Richards. You may also be interested in Gowling WLG's chapter on England & Wales in the International Bar Association (IBA) Guide on AI and Litigation.