AI and court documents: CJC backs responsible adoption, not prohibition

Technology may change how legal work is produced, but not who is responsible for it. 

That is the central message from the final report on ‘Use of AI for preparing court documents’ recently published by the Civil Justice Council. Rather than recommending blanket restrictions on AI, the CJC proposes a functional, risk-based framework focused on what the technology is actually being used to do and the risk that use creates.

Regulation not prohibition

The CJC recommends that responsibility for guidance to the legal profession should remain primarily with professional regulators such as the Solicitors Regulation Authority and Bar Standards Board. It would supplement that framework with a new Practice Direction consolidating the courts’ existing approach to accuracy, verification, responsibility and sanctions for misuse of AI.

Importantly, the report does not recommend that lawyers should generally have to declare when AI has been used to prepare statements of case, skeleton arguments or other advocacy documents. Nor does it recommend requiring lawyers to identify to the court every AI tool used or to seek the court’s permission before using AI.

Instead, lawyers remain responsible for the accuracy and reliability of the documents they produce. Where AI is used, the report recommends looking at its function and the resulting risk rather than attempting to regulate particular categories of technology.

AI technology is developing too quickly for regulation based around particular tools or labels to remain effective for long. The CJC, therefore, recommends against categorising AI as “generative”, “non-generative”, “administrative”, “analytical” or “substantive”. The proposed framework instead asks whether AI is capable of generating or influencing substantive content submitted to the court and, if so, what risks that creates for the integrity of proceedings and the administration of justice.

Witness statements are different

The report takes a deliberately stricter approach to factual witness evidence.

The CJC recommends amending the Civil Procedure Rules so that a witness of fact must specifically declare that any AI used in preparing their witness statement has not generated or influenced its substantive content. This is designed to preserve the fundamental principle that a witness statement should contain the witness’s own evidence in their own words.

Again, the distinction is functional. The proposed declaration is not designed to prohibit a particular AI product or technology. It is directed at whether AI has affected the substance of the evidence.

That is an important distinction for law firms developing their own AI policies. A policy which simply says that a particular tool can or cannot be used may increasingly be too blunt. What matters is what the lawyer is asking the technology to do, the material on which it is operating and how its output is ultimately being used.

Transparency for expert evidence

The CJC also recommends a specific declaration where AI has generated or influenced the substantive content of an expert’s report. The declaration should contain enough information to allow the opposing party and the court to scrutinise the evidence effectively. Where necessary, this should include disclosure of the AI tool used. 

This approach will allow scrutiny of the expert’s methodology, the appropriateness of using AI for that purpose and the resulting substantive content. This reflects a wider principle that is becoming increasingly important as AI tools become embedded in professional work: the greater the significance of AI to the substance of an output, the greater the need to understand, test and, where appropriate, explain its role.

A pressing issue for litigants-in-person

The use of AI by litigants-in-person was outside the working group’s original terms of reference. Nevertheless, the report identifies as a significant gap the absence of clear and accessible guidance.

The CJC recommends urgent work with the Civil Procedure Rule Committee on how litigants-in-person should use AI in civil proceedings. This will require a careful balance. AI has the potential to improve access to legal information and assistance, but inaccurate or fabricated content can also harm litigants, opposing parties and the court.

What does this mean in practice?

The report reinforces the message from the SRA’s warning notice on the misuse of AI: responsible adoption requires governance, verification and professional judgement, not a retreat from the technology.

The CJC’s recommendations recognise that AI is already being used in legal proceedings. They seek to create a framework capable of accommodating further technological change.

The proposed approach is neither unrestricted adoption nor blanket prohibition. Rather, it is one of accountable use, proportionate safeguards and continuing human responsibility.