WRC Guidance on the Use of AI in Legal Submissions

According to new research from AI Ireland, 74% of Irish businesses are now actively using artificial intelligence (“AI”), with Microsoft’s new AI Diffusion Report ranking Ireland 4^th in the world for AI adoption. AI now touches nearly every part of our daily lives, so it was inevitable that artificial intelligence would become a significant consideration within employment law and within the legal sphere generally.

The Workplace Relations Commission (the “WRC”) has published timely guidance on the responsible use of generative AI in preparing WRC submissions. This follows a recent case that highlighted and warned against the risks of misusing AI in legal proceedings.

Fernando Oliveira v Ryanair DAC

In the highly publicised case, Fernando Oliveira v Ryanair DAC (ADJ-00055225), a former flight attendant brought claims of discrimination on the grounds of race and family status, along with allegations of victimisation, harassment, sexual harassment, and procedural unfairness. The Complainant had sought €170,000 in compensation from his employer, who categorically denied the allegations and raised concerns about the Complainant’s submissions which they suspected had been generated with the use of AI.

In her decision, the Adjudication Officer noted that parties in proceedings have an obligation to ensure that the content of their submissions is relevant and accurate. She criticised the suspected use of AI in preparing the submissions, noting that they were “rife with citations that were not relevant, mis-quoted and in many instances, non-existent.” The Adjudication Officer went so far as to describe the reliance on “phantom citations” as “egregious and an abuse of process,” and took the opportunity to remind the parties that submissions must be relevant, accurate and not misleading.

In another recent WRC hearing, a HR professional admitted to using “AI assistance” to prepare legal submissions for their own WRC unfair dismissals case. The Adjudication Officer described the cases referred to in their submissions as being “hallucinations”, a term which is used to describe incorrect or misleading results that AI models generate. The employer’s legal representatives stated that the Complainant’s submissions referenced some real cases, misquoted others, and there were references to past WRC rulings which simply could not be found at all. The matter was ultimately adjourned until the end of February to allow for new submissions. It will be interesting to see whether the Adjudication Officer comments specifically on the use of AI in preparing the legal submissions and on the unreliable citations relied upon, in reaching the decision.

WRC Guidance on AI Use

Published on 30 October 2025, the WRC’s Guidance sets out how parties may responsibly use AI tools when preparing written submissions or documents for hearings. The key message is clear: parties remain fully responsible for the content of their submissions, even if AI tools were used in their preparation. Incorrect or misleading material may negatively impact a case.

The WRC highlights several risks associated with AI use, including:

• Inaccurate or fictional references
• Misleading or incorrect legal content
• Data protection and privacy concerns

The WRC recommends best practices such as:

• Double-checking all legal references to ensure they are accurate and relevant under Irish law
• Submitting only material that the party fully understands and can explain if questioned
• Avoiding the inclusion of sensitive personal data in public AI tools
• Recognising that AI cannot provide legal strategy or assess case strengths

The WRC has warned that the misuse of AI could undermine arguments, delay hearings, require corrections, and in serious cases, damage parties’ credibility. Whilst the WRC has not said that parties must disclose that AI tools were used to assist in the preparation of submissions, they have highlighted that the disclosure “promotes transparency and helps the WRC understand how your submission was prepared”

The WRC guidance makes clear that AI cannot substitute a party’s understanding of their case or the applicable law, and Adjudication Officers retain discretion to reject material that is superfluous, incorrect, or misleading due to AI generation.

The Law Society of Ireland

In parallel, the Law Society of Ireland has issued its own Guidance for solicitors on the ethical and professional use of generative AI. This guidance highlights both the potential efficiencies AI can offer in drafting, and the limitations of its use. Solicitors are reminded that professional judgement and human oversight cannot, and should not, be replaced by technology.

Conclusion

AI tools can be valuable in supporting preparation for hearings and assisting both solicitors and laypersons bringing claims before the WRC with drafting submissions and other documentation. However, we are reminded that these tools must be used responsibly, with full awareness of their limitations. It is clear from the emerging case law and latest guidance that human judgement, accuracy, and professional ethical standards remain paramount.

How can we help

If you have any queries or concerns, or would like to discuss the above in further detail, please feel free to contact Richard Lee, Head of Employment or Leanne Hill, Solicitor, in the Employment & Benefits Department of BHSM LLP on rlee@bhsm.ie / lhill@bhsm.ie.

This article is for general information purposes.  Legal advice must be obtained for individual circumstances.  Whilst every effort has been made to ensure the accuracy of this article, no liability is accepted by the author for any inaccuracies.