What Does High Court Practice Direction HC 142 Mean for the Use of AI in Litigation?

David Curran, Partner

Claudia Clifford, Legal Intern

 

The High Court's new Practice Direction HC 142, which takes effect on 1 September 2026, does not prohibit the use of Generative AI ("GenAI") in litigation. However, it places clear responsibility on lawyers, litigants, witnesses and experts to independently verify any AI-assisted content before it is used in court proceedings.  

Key Takeaway: Generative AI can be used in High Court proceedings, but users remain personally responsible for the accuracy, reliability and integrity of all material placed before the Court.  

 

What Has Changed? 

On 29 July 2026, the President of the High Court, Mr Justice David Barniville, issued Practice Direction HC 142, the first detailed standalone framework governing the use of Generative AI in High Court proceedings.  

The Practice Direction builds on the principles identified by Costello J in Guerin v O'Doherty [1] and seeks to establish clear expectations for the responsible use of AI in litigation. Its primary concern is the risk of AI-generated inaccuracies, including fabricated authorities, incorrect legal propositions and manipulated evidence.  

The Practice Direction is not intended to prevent the use of AI. Rather, it seeks to ensure that any use of AI is undertaken responsibly and subject to appropriate human oversight.  

 

Who Is Affected? 

From 1 September 2026, the Practice Direction applies to all persons who introduce documents to the High Court, including:

• Solicitors 

• Barristers 

• Litigants in person 

• Witnesses 

• Expert witnesses 

• Third parties involved in proceedings 

It applies to all documents prepared, produced or generated within proceedings after 1 September 2026, regardless of whether the proceedings themselves were commenced before or after that date. This means that documents still being drafted, including pleadings, affidavits, witness statements and expert reports, may be subject to the new requirements even where proceedings are already well advanced.  

 

What Are the Main Requirements?

The High Court expressly recognises that AI can improve efficiency, reduce costs and enhance access to justice. However, the central principle of the Practice Direction is that responsibility for court documents remains with the human author.  

The use of AI does not reduce: 

• A solicitor's or barrister's professional obligations; 

• A witness's responsibility for their evidence; 

• An expert's obligation to provide an independent opinion; or 

• A litigant's responsibility for documents filed with the Court.

 

Independent Verification Is Mandatory 

Any material generated wholly or partly through AI assistance must be independently verified by a human before it is relied upon in proceedings. Verification by another AI system is not sufficient.  

The verification obligation goes beyond checking citations. Users must ensure that: 

• Cases, legislation and authorities actually exist; 

• Authorities support the propositions attributed to them; 

• Content is current, accurate and complete; and 

• Material is relevant to the issue under consideration.  

The Practice Direction also requires particular caution regarding images, audio and video evidence. Fabricated, manipulated or deepfake materials must never be presented to the Court.  

From a practical perspective, firms should consider maintaining a record of AI use and verification steps undertaken so that an audit trail is available if questions arise later.  

 

Do Parties Have to Disclose Their Use of AI? 

One of the more notable aspects of HC 142 is that it does not impose a general obligation to disclose AI use in ordinary court documents.  

Parties are not required to routinely state that AI was used in preparing pleadings, submissions or correspondence. However, they must be capable of explaining: 

• Whether AI was used; 

• The nature of that use; and 

• The extent of that use, 

if requested by the Court. 

Specific disclosure and declaration requirements do apply to certain categories of documents.

 

What Rules Apply to Particular Court Documents? 

Pleadings 

For pleadings, parties must ensure that: 

• Facts are accurate; 

• Legal foundations for relief sought are sound; 

• Authorities exist and support the propositions advanced; and 

• Any chronology relied upon is correct.  

 

Written Submissions 

All authorities, quotations, references and citations included in written submissions must be independently checked against authoritative sources.  

Solicitors receiving submissions from counsel may increasingly seek confirmation regarding whether AI has been used and, if so, the extent of that use.  

 

Affidavits and Witness Statements 

The Practice Direction adopts a particularly cautious approach to witness evidence. 

Every affidavit and witness statement prepared after 1 September 2026 must contain a prescribed declaration confirming that AI has not been used to generate or alter the substantive content of the evidence.  

The fundamental principle is that evidence must reflect the witness's own knowledge, recollection and evidence. Lawyers should therefore avoid using AI to reshape, embellish, strengthen, dilute or otherwise alter how a witness expresses their evidence.  

Administrative functions such as formatting and spellchecking remain permissible, provided AI has not been used to generate substantive evidence.  

 

Expert Reports 

Expert reports prepared after 1 September 2026 must also contain a prescribed declaration.  

While AI may be used in preparing expert reports, any material use of AI in drafting, analysing, interpreting or generating substantive content may need to be disclosed where necessary to assess the reliability and independence of the expert opinion.  

Solicitors instructing experts should ensure that experts understand these requirements from the outset and acknowledge their obligations before commencing work.  

 

What Happens If an AI-Generated Error Is Found? 

Where a party discovers an AI-generated inaccuracy in its own documents, the Court and the other parties must be notified without delay. The responsible party may also be required to bear the costs of correcting the error. 

If one party identifies an inaccuracy in another party's document, the issue should first be raised directly with that party. If the matter is not appropriately addressed, it may then be brought before the Court.  

It is likely that scrutiny of AI use in litigation will increase and that parties will examine documents carefully where AI involvement is suspected.  

 

What Are the Risks Around Privilege and Confidentiality? 

The Practice Direction highlights the potential risks associated with entering confidential, privileged or restricted information into AI systems. Information uploaded to certain AI tools may be retained, processed or reused in ways that are not immediately visible to users.  

Lawyers should carefully assess AI platforms before use and ensure that confidentiality, legal privilege and court-ordered restrictions are not compromised. Documents subject to privilege, court orders or confidentiality undertakings require particular caution.  

 

What Are the Consequences of Non-Compliance? 

Failure to comply with Practice Direction may have significant consequences, including: 

• Rejection of court documents; 

• Striking out of claims or defences; 

• Adverse costs orders; 

• Personal costs orders against lawyers; and 

• Referral to the Legal Services Regulatory Authority.  

These obligations apply equally to represented parties and litigants in person. 

 

What Should Lawyers and Litigants Do Now? 

The key message from the High Court is straightforward: AI has a legitimate role in modern litigation, but responsibility remains with those who use it.  

Practical steps include: 

• Reviewing internal AI policies; 

• Implementing verification procedures for all AI-assisted work; 

• Maintaining records of significant AI use; 

• Updating witness and expert briefing processes; 

• Ensuring clients understand the new declaration requirements; and 

• Carefully assessing confidentiality and privilege risks before inputting information into AI systems.  

The Practice Direction provides a significant framework for the responsible use of Generative AI in Irish litigation and is likely to influence the development of further guidance in this rapidly evolving area.  

 

Contact Holmes 

If you have questions about the impact of Practice Direction HC 142 or the use of AI in litigation, contact our Litigation & Dispute Resolution team or David Curran for advice on your specific circumstances.

 

[1] [2026] IECA 48.

 

Dislaimer of Liability: This insight article is provided for general information purposes only and should not be relied upon as legal advice. Readers should seek specific legal advice tailored to their individual circumstances before taking or refraining from any action. Holmes O'Malley Sexton LLP, its partners, employees, agents, and affiliated entities disclaim all liability, to the fullest extent permitted by law, for any loss or damage arising directly or indirectly from reliance on the information contained in this publication.