What is the role of Gen AI in litigation?

This series is written for in-house legal teams and business leaders who want to take a proactive and practical approach. Throughout the series, we examine the direction of travel from those shaping litigation practice. We will look at what the Civil Justice Council, legal regulators such as the Law Society and the Courts and Tribunals Judiciary are saying about the use of AI and the risks it creates particularly around privilege and evidence. These developments are beginning to set expectations around governance, transparency and professional responsibility. Understanding that emerging guidance, and the steps that can be taken now, will be critical for organisations seeking to ensure their adoption of AI takes advantages of the opportunities while mitigating the risk.

In Part 1 we review the CJC’s interim report, which sets out its initial thinking on whether rules are required to govern the use of AI by legal representatives for the preparation of court documents. The proposals seek to strike a balance between using the latest technology to enhance access to justice and improve efficiency while maintaining confidence in the rule of law. Key risks include hallucinations/generation of fictitious case citations and concerns around the authenticity of evidence.

What is the role of Gen AI in litigation?

The Civil Justice Council (CJC) has published an interim report setting out its initial thinking on whether rules are required to govern the use of AI by legal representatives for the preparation of court documents.

The proposals seek to strike a balance between using the latest technology to enhance access to justice and improve efficiency while maintaining confidence in the rule of law. Key risks include hallucinations/generation of fictitious case citations and concerns around the authenticity of evidence.

This is a challenge facing all jurisdictions and the CJC take a brief look at practice and procedure in courts around the world, concluding that there is no international consensus on these issues.

In summary, the CJC proposes that legal representatives should make a declaration on the use of AI where it has been used to generate evidence on which the court is being asked to rely. No declaration would be required for other uses of AI, which are administrative in nature. The CJC is consulting on its proposals. The deadline for responses is 14 April 2026.

In this article, we consider the proposals, how these fit with existing duties and what this means for the future conduct of litigation.

What are the proposals?

Statements of case, skeleton arguments and other advocacy documents

The CJC recognises that these documents are prepared by legal representatives who are subject to professional regulation. Therefore, the CJC proposes that provided the statement/document bears the name of the legal representative who is taking professional responsibility for it, there is no need for any further rules relating to the use of AI in the preparation of these documents.

Witness statements

There are different proposals for trial and non-trial witness statements.

For trial witness statements, the existing rules make clear that they should always be in the witness’s own words. On that basis, the CJC notes that it is difficult to see how AI could be used in their preparation other than for non-text/administrative purposes such as spell checking or transcription. Consequently, the CJC proposes a rule change (for both PD57AC and CPR 32) to require a declaration that AI has not been used for the purposes of generating the statement. 

The CJC also raises questions on the use of AI for translation of witness statements which have been created in a foreign language. Currently, when a witness statement is filed in a foreign language, the rules require that a translation be provided and that a translator certify the translation is accurate. The CJC asks whether a rule is needed to allow use of AI by human translators, but notes that may not be necessary given that the translator is taking responsibility for it. The CJC also asks whether the rules should be amended to allow publicly available machine translations or whether use of such tools should be restricted to legal representatives. 

Witness statements which are not for trial (usually made in support of interim applications) are often drafted by legal representatives, so the CJC suggests the same approach is adopted as for statements of case. Therefore, provided the document bears the name of the legal representative who is taking professional responsibility for it, then there is no need for any further rules relating to AI.

Disclosure 

The use of AI in connection with disclosure is already well established and the CJC notes that anecdotal evidence suggests that parties are co-operating in their use of AI to conduct disclosure exercises. On that basis, the CJC does not think there is a pressing need to amend disclosure statements/lists to include a section addressing the extent to which AI tools have been used.

The Law Society has also published practical guidance on the use of Generative AI in legal disclosure.

Expert reports

The Bond Solon Expert Witness Survey 2025 reports that 20% of experts had used AI in their role as an expert witness. Bond Solon cites the American case of Kolhls v Elison No 24 in which an expert witness had used AI to draft their report and accidentally included misinformation and therefore suggest that further guidance is required. 

The CJC recognise that while experts are subject to specific duties under the Civil Procedure Rules, there will be circumstances in which a failure to disclose the use of AI could lead to erroneous evidence being put before the court. Therefore, the CJC proposes that the expert explain what use of AI has been made (other than administrative) and that the expert disclose the AI tools used. 

While any rules changes are confirmed, experts may wish to consult the recent Academy of Experts guidance on the use of AI, published in January 2026.

Existing duties 

These proposed new rules would sit alongside existing obligations for legal representatives to act with honesty and integrity and not to mislead the court or advance propositions which are not properly arguable. 

The recent guidance for judges on the use of AI (published in October 2025) makes clear that “all legal representatives are responsible for the material they put before the court/tribunal and have a professional obligation to ensure it is accurate and appropriate.”

There have already been numerous cases before the court involving hallucinations where legal representatives have been sanctioned under existing duties for including references to cases or legislation which do not exist. A recent example is the case of Ayinde v. London Borough of Haringey [2025] in which legal representatives cited five fictional cases. In her judgment in that case, Dame Victoria Sharp noted that there are a range of powers open to the court in such a situation, including costs penalties, strike out, referral to a regulator, contempt proceedings or referral to the police. 

Use of AI by litigants in person

The proposed rules apply to legal representatives, but not litigants in person (LiPs). The CJC recognises that AI chatbots are increasingly being used by unrepresented litigants and this is an area that would benefit from further work.

The CJC notes that AI has the potential to assist with access to justice, and suggests that in future, LiPs may also be required to declare whether AI has been used. For now, the key source of guidance applicable to LiPs is set out in Judicial Guidance on AI published in October 2025, in which it is said: 

“If it appears an AI chatbot may have been used to prepare submissions or other documents, it is appropriate to inquire about this, ask what checks for accuracy have been undertaken and inform the litigant that they are responsible for what they put to the court.” 

Approach to use of AI in other jurisdictions

Guidance on the use of AI in court proceedings has been issued in a number of jurisdictions across the world, including the United States and Hong Kong but the CJC report notes that there does not yet appear to be an international consensus. 

The CJC notes that in Singapore, there is no prohibition on the use of AI to prepare court documents and no requirement to make a declaration upfront about the use of AI. That contrasts with the position in some jurisdictions in the US, where declarations regarding the use of AI are required. A number of courts require attorneys to file a notice outlining their use of AI and certifying that all AI generated content, especially legal citations have been verified using traditional approaches.

The Chartered Institute of Arbitrators has also issued its own guidelines on the use of AI, with the most recent update published in September 2025. This includes a template agreement which can be used by the parties to an arbitration on the use of AI.

Privilege

The impact of AI on legal professional privilege is a hot topic in the industry. That issue is not covered by the CJC report, but we are starting to see the courts grappling with it. The starting point is that the existence of new technology does not immediately change the way the law works and the courts will apply the existing law of privilege when determining whether a particular document is protected. 

There are a range of issues to consider, particularly in relation to the loss or waiver of privilege in existing documents, as well as the creation of new documents which do not benefit from the protection afforded by legal professional privilege and may therefore become disclosable to counterparties, courts and regulators. 

Next steps 

The CJC’s report and consultation recognises the transformative potential of AI in legal proceedings and seeks to balance efficiency with the need for safeguards. The proposed rules do not represent a substantive change, but serve as a framework for responsible AI adoption, reinforcing the core values of transparency, accountability, and justice in the courts of England and Wales.

The consultation closes on 14 April 2026. There is no indication yet as to when the final report is likely to be published and when any new rules are likely to come into force. There may also be further work to consider issues outside the scope of this CJC report, such as the use of AI by LiPs.