In February 2026, the Civil Justice Council (CJC) issued a consultation, together with an interim report, on the use of AI for preparing court documents. Following initial analysis of consultation responses, the CJC issued an update in June 2026, summarising that for professional legal drafting – pleadings, skeleton arguments, advocacy documents – there was broad agreement that no additional AI-specific rules are needed. The principal area identified for ongoing consideration was the use of AI in the preparation of witness statements.
The CJC has now issued a final report and makes a series of recommendations, aimed at promoting the safe and responsible use of AI in the preparation of court documents and at adopting a functional, risk-based approach to AI use.
We examine some of the key recommendations.
Responsibility rests with legal representatives and professional regulators, but new Practice Direction recommended
The professional responsibility of legal representatives has been reiterated in various cases before the courts where AI misuse has come to light and guidance has been given to the profession by bodies including the Solicitors Regulation Authority and the Bar Standards Board. The report concludes that the starting point for the regulation of AI use in the preparation of court documents should remain with legal representatives and their professional regulators and the latter should remain responsible for the provision of guidance on appropriate use of AI. This allows more detail to be provided and updates to be made more easily, with continuing training provided alongside.
However, the CJC has recommended that professional regulation should be supplemented by a regularly updated Practice Direction to codify and consolidate the existing approach of courts and tribunals to the use (and sanctions for misuse) of AI by legal representatives and litigants in person in civil proceedings.
Urgent need for guidance for litigants in person (LIPs)
The scope of the consultation did not include the use of AI by litigants in person. However, due to the strength of feedback in consultation responses, and the fact that the lack of guidance represents a significant gap in the current framework, a recommendation has been made that the CJC, together with the Civil Procedure Rules Committee (CPRC), should urgently consider the approach that should be taken to the use of AI by LIPs in civil proceedings. This would include guidance on the use of publicly available AI tools by LIPs.
Further, the recommended Practice Direction would apply to LIPs, as well as legal representatives.
Declaration required in witness statements as to AI use
This was the most contentious area within the consultation, in light of differing views in consultation responses as to whether additional safeguards or disclosure requirements are needed.
The CJC has recommended that:
"The CPR should be amended to require a witness of fact who makes a witness statement to make a specific declaration that AI used in its preparation has not generated or influenced its substantive content." (emphasis added)
The aim of this recommendation is to secure the integrity of the witness' evidence and to ensure that it is given in their own words.
Guidance has been given in the report as to the meaning of the highlighted terms:
- 'Generate substantive content' – "refers to the production or modification of substantive content by an AI tool, whether intentionally or unintentionally".
- 'Influence substantive content' – "refers to the use of an AI tool which, without itself producing such content, affects its selection, analysis, development or evaluation".
Whilst the concept of AI generating or modifying substantive content in a witness statement can be readily understood, the question of whether AI has 'influenced' substantive content seems more problematic given the potential wide scope of this term and the possibility of disagreement as to whether a particular use of AI has influenced content or not.
The CJC recommends that the CPR should identify activities capable of 'influencing'. The concept as currently explained could encompass the work in preparation for witness proofing and could also conceivably catch sensible uses such as analysing interview transcripts or ordering raw evidence into a coherent chronological order. Therefore, careful explanation will be required to ensure that the proposed scope is clear and not wider than intended. This may otherwise become an area for future uncertainty and argument, and may deter parties from using AI where it is efficient and cost-effective to do so.
It would be for the witness to make the declaration, albeit in order to do so, they would need sufficient information from their lawyers regarding the preparation of the statement. For this reason and to safeguard against any future challenge, it will therefore be important to monitor for compliance, keep underlying records and be able to explain how any AI tool was utilised.
Also of interest is the statement in the report that use of AI to generate or influence substantive content "can be viewed as analogous to witness coaching". This replicates the Court of Appeal's warning in criminal case, R v FGD [2026] EWCA Crim 918, that coaching witnesses on the substance of their evidence by any means, including AI, is prohibited, and witnesses should be "firmly discouraged" from using AI to prepare.
If a witness statement is being made by a legal representative (who is not making the statement as a witness of fact), responsibility as to its accuracy and reliability remains with that individual. The functional and risk-based approach to AI use should be applied (and should be made clear in the new Practice Direction or other CPR amendments) and no declaration will be required.
Experts to make declaration as to AI use
The report acknowledged that increasing numbers of expert witnesses are using AI in their role as experts and noted that there was strong support for experts to disclose and explain any substantive AI use. The CJC has recommended that experts should also make a specific declaration concerning AI use "where such use generates or influences the substantive content of the expert's report". Such declaration is to secure transparency and should be accompanied by sufficient information to enable effective scrutiny and testing of an expert's evidence, including the methodology used, the appropriateness of AI use for that purpose and its substantive content. Where necessary to enable such scrutiny, the information required will include disclosure of the AI tool used.
Revised guidance is also recommended to embed good practice in AI use by experts in civil claims, including making clear that three questions need to be considered by experts:
- First, whether their use of AI has generated or influenced the substance of their evidence. If so, the declaration needs to be made.
- Secondly, and if the declaration should be made, what information should be disclosed to enable the opposing party and court to scrutinise their methodology and analysis effectively?
- Thirdly, whether it is necessary to disclose the AI tool used as part of the information disclosed.
The importance of transparency and responsibility around experts' use of AI has recently been highlighted by the courts. Although a family case, Cornwall Council v M [2026] EWFC 306 is a stark illustration of the issues that can arise where AI is used in the preparation of expert evidence, and of the court's expectation that such use is disclosed and capable of scrutiny. The expert explained that she had used AI as an administrative tool and had reviewed, amended and verified all content. However, it transpired that AI use was more extensive, included errors and was submitted complete with AI commentary (for example, "This version is about half the length, avoids advocacy ... and is more likely to withstand cross-examination because ..."). For this, and other reasons, the court did not feel able to place any significant weight on her evidence.
No other declarations recommended
As indicated in the June 2026 update, no new rules have been proposed requiring legal representatives to make a declaration concerning AI use in the preparation of statements of case, skeleton arguments or other forms of advocacy. This is in line with the position that legal representatives remain responsible for the content of these documents under existing professional obligations.
Neither were reforms suggested in respect of AI use in disclosure. This is subject to any recommendations made by the Disclosure Review Working Group.
Finally, rather than providing a technical definition of "AI", it is recommended that the CPR should define the scope of AI use functionally, asking whether the activities are capable of generating substantive content or influencing the nature and development of content submitted in civil proceedings. This is in order to avoid disputes as to whether a technology falls within scope of a definition and to avoid it becoming outdated due to rapid technological change. This fits in with the proposed risk-based approach to AI use.
Next steps
These are recommendations, which are not yet in force – the CPRC will now consider them for implementation. The CJC will review the extent to which the recommendations have been taken forward in its next annual review. Watch this space for further developments.
