Civil Justice Council Consults on Rules for AI Use in Court Documents

The Civil Justice Council has published an interim report and consultation paper examining whether formal court rules are necessary to govern the use of Artificial Intelligence (AI) by legal representatives and experts when preparing court documents. Led by a working group made up of several established members of the legal industry, the document seeks to establish a clear framework within the civil justice system. 

The primary objective is to allow legal practitioners to leverage modern technology to increase efficiency and lower costs while maintaining strict compliance with established ethical standards and professional duties. The report also explicitly distinguishes between administrative or technical applications of technology, such as spell-checking, transcription, formatting, and accessibility tools, and generative AI systems that produce substantive legal content, reasoning, or evidence.

Key proposals by document type

The Civil Justice Council outlines proposals based on the document type and whether it constitutes legal advocacy or factual evidence:

Statements of case, skeleton arguments and advocacy documents

The CJC proposes that no new court rules or specific AI declarations are required for statements of case, skeleton arguments, chronologies, or lists of issues. These materials represent exercises in legal advocacy drafted by regulated legal professionals who are subject to existing professional conduct rules. 

As long as the document explicitly identifies the legal representative taking professional responsibility for its contents, existing duties to the court are sufficient. The consultation does, however, invite views on an alternative option: requiring a mandatory declaration that details whether and how AI was used during preparation.

Witness statements

For trial witness statements, including those governed by Practice Direction 57AC and Civil Procedure Rules (CPR) Part 32, the CJC proposes amending court rules to require a mandatory declaration from the relevant legal representative confirming that generative AI was not used to draft or modify the substantive content. This restriction prohibits AI from altering, embellishing, strengthening, diluting, or rephrasing witness evidence, guaranteeing that statements remain strictly in the witness's own words. 

For non-trial witness statements, such as those used in interim or procedural applications, the CJC suggests no additional AI rules are necessary, provided the named legal representative takes professional responsibility. On the topic of foreign language statements, the paper considers whether court rules should permit human translators or legal representatives to use machine translation tools, provided the specific tool is identified and the translation is certified for accuracy.

Expert reports

Unlike fact witnesses, expert witnesses may legitimately use AI tools provided that their use is fully transparent. The CJC proposes amending Practice Direction 35 (PD35) to require experts to declare and explain any use of AI in preparing their reports, excluding routine administrative uses like transcription, and to explicitly identify the specific AI tools deployed.

Disclosure

The CJC concludes that there is no immediate need to modify disclosure rules to require parties to declare AI or Technology Assisted Review (TAR) software in disclosure lists. Current practice demonstrates that litigating parties are successfully collaborating on technology use during disclosure without requiring judicial intervention.

Professional obligations and AI risks

The report emphasises that legal representatives remain strictly accountable for all material submitted to the court. Regulatory frameworks, including the Solicitors Regulation Authority Code of Conduct and the Bar Standards Board Conduct Rules, require practitioners to act with honesty and integrity, independently verify legal authorities, and avoid misleading the court.

The consultation highlights specific risks associated with generative AI in litigation. A primary concern is hallucination, where large language models fabricate non-existent case citations, statutory provisions, or legal texts. Submitting fake citations breaches professional duties and can lead to judicial admonition, wasted costs orders, or regulatory penalties. Additional risks include embedded bias within AI training models and manipulated evidence, such as deepfakes or hidden white text prompts designed to alter search engine and model outputs.

Wider context and litigants in person

While the consultation focuses primarily on procedural rules for regulated legal representatives, the CJC notes the increasing use of AI chatbots by unrepresented litigants. Because litigants in person lack formal legal regulation and may struggle to verify AI outputs, the CJC indicates that separate work may be required in the future to address AI use by unrepresented parties without impeding access to justice.

Implications for TrialView and the future of digital litigation

The CJC interim report provides essential clarity for legal technology platforms, confirming that AI tools designed for workflow efficiency, data analysis, and court compliance remain fully aligned with civil justice principles. 

The proposed regulatory framework draws a clear distinction between unverified, generative AI drafting and structured, administrative, or analytical AI assistance.

This boundary reinforces the core architecture of platforms like TrialView. TrialView assists legal teams through document organisation, automated hyperlinking, OCR, timeline creation, evidence analysis, and compliant court bundle generation. These capabilities fall squarely under the CJC's permitted category of analytical and administrative technology. By maintaining human oversight over legal advocacy and keeping fact witness evidence strictly in the witness's own words, platforms like TrialView enable legal professionals to process complex evidence efficiently while fully satisfying their professional duties to the court.

Consultation questions

The CJC invites responses from legal professionals, the judiciary, and technology stakeholders on the following key questions raised in the report:

• Should guidance on AI remain the responsibility of professional bodies, or should formal court rules be introduced?
• Do you agree that statements of case and skeleton arguments require no formal AI declaration if a legal representative takes professional responsibility?
• Should trial witness statements require a mandatory declaration confirming that generative AI was not used to produce or modify the text?
• Should expert witnesses be required to declare and detail their use of AI tools in court reports?
• Is the term artificial intelligence sufficiently defined for the purpose of civil procedure rules?