Irish Legal News has seen the first formal guidelines for Irish judges on the “responsible use of generative AI”, which were drawn up by a judicial committee last summer and closely mirror guidelines in England and Wales.

The document, which will be published today, contains basic explanations of how generative AI works and highlights ways in which it can be used by judges as well as by lawyers and litigants-in-person.

Judicial use of GenAI is coming under increasing scrutiny across the world. In the US, a gaffe by two federal judges who issued court orders riddled with AI-generated errors recently led to calls from the chairman of the US Senate judiciary committee for binding judicial guidelines.

Guidance for judges in England and Wales was originally published last April and updated last week.

Irish judges have received training on AI and will next year take part in further training events “in collaboration with other common law jurisdictions” which will focus on AI and its impact on the judiciary.

The first formal guidance for Irish judges was drawn up last year by a judicial committee established by the Chief Justice Donal O’Donnell outside of the Judicial Council.

The 11-page document outlines that members of the judiciary should ensure that their use of AI tools is “consistent with the overriding obligation to ensure the integrity and reliability of the legal research and analysis which underpins judicial decision making”.

Any use of generative AI tools “must be consistent with the judiciary’s overarching obligation to protect the independence and integrity of the administration of justice and the protection of fundamental rights”, it adds.

The guidelines state clearly: “Judges remain responsible for their decisions.”

Where judicial support staff, including judicial assistants, make use of generative AI tools, this must be disclosed to their supervising judge, the guidelines say, in order to “ensure that judges are alerted to the need for particular vigilance and scrutiny” of their work.

Ethical issues and concerns around copyright and data protection are also highlighted throughout the guidance, with judges warned of “gender and racial biases” in AI as well as the risk of “unintentional disclosure of private, confidential, suppressed or privileged information”.

The guidelines also address the use of generative AI by others, including solicitors, barristers and lay litigants, and the challenges posed by ‘deepfake’ technology.

The Workplace Relations Commission recently published new guidance for litigants on the use of generative AI tools following a case where GenAI was blamed for the inclusion of “phantom citations” in submissions from a complainant without legal representation.

The guidelines for judges highlight “red flags” which could indicate the use of generative AI in submissions, including the use of American spellings and references to foreign cases.

Neither legal representatives nor lay litigants need to disclose the use of generative AI, the guidelines state, but judges may quiz them where concerns arise.

“Provided AI is used responsibly, and the output is appropriately tested and verified there is no reason why a legal representative ought to refer to its use, but this depends on context,” the document says.

“Judges may request confirmation that legal professionals have independently verified the accuracy of any research or case citations generated by a GenAI chatbot if a judge has any concerns regarding the materials provided to the court.”

However, lay litigants or McKenzie friends “will often not have the skills or training to independently verify the legal information provided to them by a GenAI chatbot”, it notes.

“If it appears that a GenAI chatbot may have been used to prepare submissions or other documents by an unrepresented party, a judge may need to inquire whether a GenAI chatbot has been used and in what manner, and to ask what checks for accuracy have been undertaken (if any).”

A spokesperson for the Judicial Council, which did not produce the guideline document, said it “should be seen as an evolving document which will be updated on an ongoing basis to reflect the latest developments in GenAI technology and its applications”.

Dr Brian Barry, associate professor in the School of Law at Trinity College Dublin, told Irish Legal News that the Irish guidelines draw extensively from the guidelines in England and Wales, which he said reflected “a sensible approach”.

He said: “A key factor is maintaining this as a living document given the rapid pace of developments. It’s also encouraging that judicial training on AI is high on the Judicial Council’s agenda.

“These guidelines and judicial training could serve as a springboard for the Irish judiciary to look to other jurisdictions where AI is already enhancing the court experience — using AI technology to provide clearer public information, streamline case management, and guide litigants through the court process.”

The Judicial Council spokesperson said: “The Judicial Studies Committee, on behalf of the Judicial Council, is responsible for overseeing the education and training of the Irish judiciary in accordance with section 17 of the Judicial Council Act 2019.

“Judicial education in the area of artificial intelligence is delivered through a range of initiatives including programmes run by national and international agencies.

“Over the past number of years, members of the judiciary have participated in programmes and workshops organised by EU legal training institutions focusing on various aspects of AI and digitalisation.

“Training provided to the judiciary has included the ethical and practical implications of AI in legal decision-making, judgment writing, and the evaluation of evidence.

“A series of training events is being planned for 2026 in collaboration with other common law jurisdictions to explore AI and its impact on the judiciary.”

The guidelines for judges in relation to the use of AI “will be published on our website shortly”, the spokesperson added.