Generative AI in Irish Courts: High Court Practice Direction HC142
Artificial intelligence is everywhere right now. Businesses and individuals alike are finding new and cost-effective ways to use it, whether to redesign a garden, draft correspondence or tackle everyday administrative tasks. The legal profession is no different, with firms increasingly exploring how AI can be used to improve efficiency and support legal work. But as with any new technology, it comes with risks and limitations that need to be understood.
Our litigation team has seen a noticeable increase in AI-generated correspondence and pleadings from lay litigants. While these documents can be produced in a matter of minutes, they still need to be carefully reviewed, analysed and addressed. The result is often more time spent dealing with documents that may look persuasive at first glance but do not always stand up to scrutiny.
AI-Generated Legal Content and the Risks for Litigants
The courts are already seeing the impact of AI-generated legal content. Recent examples include:
- The citation of fictitious or “hallucinated” authorities that simply do not exist, as occurred in Guerin v O’Doherty [2026] IECA 48.
- The advancement of flawed legal arguments which litigants were unable to adequately explain, as noted by the High Court in Coulston & Others v Elliott & Elliott [2024] IEHC 697.
High Court Practice Direction HC142 and Generative AI
In response, the High Court issued Practice Direction HC142, Practice Direction on the Responsible Use of Generative Artificial Intelligence in Court Documents, which came into effect on 1 September 2026.
What Happens if AI-Generated Court Documents Get It Wrong?
The Practice Direction applies to legal practitioners, parties, witnesses, experts and litigants in person. Its core message is simple: using AI does not reduce anyone’s responsibility for the accuracy of documents placed before the Court. If AI is used, its output must be independently checked and verified before it is relied upon.
The biggest change relates to evidence. Affidavits, witness statements and expert reports must now contain prescribed declarations regarding the use of Generative Artificial Intelligence in the preparation of substantive content. The aim is straightforward: witness evidence must remain the witness’s own evidence, and expert opinions must remain the expert’s own opinions.
The Practice Direction also reinforces the obligation to verify legal authorities, quotations and references, and specifically warns against the use of fabricated, manipulated or “deep fake” evidence. The consequences of getting it wrong can be significant and include adverse costs orders, rejection of court documents, the striking out of claims or defences and, for legal practitioners, referral to the Legal Services Regulatory Authority or their professional body.
The introduction of HC142 is a welcome development. It recognises that AI can be a useful tool, but it also draws an important line in the sand. AI can assist with legal work, but it cannot replace human judgment, experience and critical thinking. Put simply, if a document is signed, sworn or filed, the responsibility for it remains entirely human. In an age of AI hallucinations, the High Court has made it clear that accountability is not something that can be outsourced.
For more information please contact Sarah Bruen.
The content of this article is provided for information purposes only and does not constitute legal or other advice.
