AI in civil litigation: Use of AI for Preparing Court Documents Final Report by the Civil Justice Council

The Use of AI for Preparing Court Documents Final Report was issued on 29 October 2026 (‘the Report’) and sets out 15 recommendations for the future regulation of AI in civil proceedings, covering regulatory guidance, litigants-in-person, statements of case and advocacy, disclosure, witness statements, translations, expert evidence and the wider approach of the Civil Procedure Rules to AI. 

Overall, the Report recommends a functional, risk-based approach to AI use, with an aim to promote the safe and responsible use of AI in the preparation of court documents. The Report also takes a practical approach, proposing that a single practice direction is introduced which brings together the CJC’s recommendations and the approach developed by the courts and tribunals in case law. 

The approach taken by the Report is likely to be welcomed by a profession in which many have already embraced AI as part of their working lives. Currently the guidance on use of AI sits within developing case law and regulators’ existing guidance. Having all this in one place makes sense and brings welcomed clarity for the legal profession and litigants in person. 

Below, we summarise some of the key takeaways across the different areas: 

From regulating tools to regulating risk

CJC rejects fixed categories such as generative, non-generative, administrative, analytical, substantive or otherwise for AI tools. Even apparently administrative technology, such as transcription, can affect substantive content. Instead, the CPR should take a functional, risk-based approach, asking whether AI generates or influences substantive material and what risk that creates for the administration of justice (see paragraphs 9.19–9.21 and Recommendations 11–13).

A new Practice Direction, but not blanket disclosure of AI use

The Civil Justice Council (CJC) recommends that professional guidance should remain primarily the responsibility of regulators such as the Solicitors Regulation Authority (SRA) and The Bar Standards Board (BSB). However, this should be supplemented by a new Practice Direction (PD) to codify and consolidate the existing case law approach to AI use and sanctions for misuse while incorporating the proposed functional, risk-based approach (see paragraphs 2.2–2.3 and Recommendations 1–3). 

Nevertheless, the CJC rejects a general requirement for legal representatives to disclose which AI tool they have used. It also concludes that no new rules are needed for parties to seek permission to use AI because the court's existing case management powers, including CPR 3.1(2)(p), are sufficient (see paragraphs 9.27–9.37 and Recommendations 14–15)

Lawyers remain responsible for what goes before the court

For statements of case, skeleton arguments, and other written advocacy, the CJC recommends no new AI declaration. Instead, legal representatives should remain responsible under their existing obligations for the accuracy and reliability of documents submitted to the court (see Recommendation 4). 

The CJC considered that requiring disclosure simply because AI had been used would not necessarily improve reliability and could instead create satellite litigation about what qualifies as AI. The focus remains on the accuracy and reliability of the content, rather than the technology used to produce it (see paragraph 5.10). The CPR should be amended to clarify that, where AI is used, the specific legal professional who takes responsibility for the document is identified.

Witness statements require greater protection

For witnesses of fact, the CJC recommends amending the CPR to require the witness make a declaration that the statement has not been prepared through the use of an AI tool that has generated or influenced its substantive content. This is intended to ensure that the evidence remains the witness's own, unaffected by AI-generated or AI-influenced substantive content (see paragraphs 7.14–7.15 and Recommendation 6). The declaration focuses on the result of such use, not on the use itself or the tool that produces the result.

‘Generate substantive content’ refers to the production or modification of substantive content by an AI tool, whether intentionally or unintentionally (our emphasis added). ‘Influence substantive content’ refers to the use of an AI tool which, without itself producing such content, affects its selection, analysis, development, or evaluation. The Report notes that in either case, use of the AI tool may thus affect the substantive content, i.e. facts, evidence, opinion, reasoning, or argument contained in documents submitted in civil proceedings, including, in a witness statement, the witness’s own words. 

Greater safeguards where AI shapes expert evidence

A new declaration concerning AI use and revised guidance on AI use for experts is proposed where such use generates or influences the substantive content of the expert’s report. CJC explains that “Transparency is required to enable the expert’s evidence to be scrutinised and tested effectively by opposing parties and the court”. The declaration should provide sufficient information, including, where necessary for this purpose, disclosure of the AI tool used to the opposing party and the court, to facilitate effective scrutiny of the expert evidence, including the methodology used, the appropriateness of AI use for that purpose, and its substantive content (see paragraphs 8.10 – 8.16 and Recommendations 9 – 10).

Disclosure remains largely unchanged

The CJC recommends no immediate reform of the rules governing AI-assisted disclosure, noting that technologies such as Technology Assisted Review (TAR) are already well established and operate within PD31B and PD57AD. Subject to the ongoing Disclosure Review Working Group, the focus should remain on the integrity of the disclosure process rather than the particular tool used (see paragraphs 6.1–6.7 and Recommendation 5). 

Translations

No new substantive rules are needed. All translations must be certified by a professional human translator and be accompanied by the original text.

Urgent guidance is needed for Litigants in Person

Although AI use by litigants in person fell outside the Working Group’s original Terms of Reference, the CJC identifies the absence of clear, accessible guidance as a “pressing concern”. AI can facilitate access to justice but can also cause harm to litigants, opposing parties and the court. The CJC therefore recommends that it work urgently with the Civil Procedure Rule Committee (CPRC), informed by the judicial cross-jurisdictional AI Committee, to develop an appropriate approach (see paragraphs 4.4–4.8 and Recommendation 2).

For more information about the law, technology and practice of disclosure, contact Tom Whittaker or Stacie Bourton.