Ireland's AI Advisory Council Tells Schools and Universities: AI-Detection Tools 'Will Not' Work

In February 2025, Ireland's national AI Advisory Council — an independent expert body reporting to Government through the Department of Enterprise, Trade and Employment — published an advice paper concluding that efforts to detect AI-generated student work 'will not' succeed given the sophistication of generative AI, and recommended that schools and colleges stop relying on plagiarism/AI-detection software. It called instead for redesigning assessments, embedding AI literacy into staff continuing professional development, and documenting procurement and data-governance decisions so human oversight stays visible and defensible.

To move Irish schools and higher-education institutions away from reliance on unreliable AI-detection software and toward assessment redesign, staff AI-literacy training, and transparent, documented human oversight of AI-related decisions.

The advice was widely reported, including by the Irish Times ('Schools told to prepare for AI in class as detection methods 'will not work'', 21 Feb 2025). Several universities have since updated policy: UCC launched a university-wide misconduct register in September 2024, and the University of Limerick introduced a new academic-integrity policy addressing generative AI for 2025/26.

The Council's caution has since been borne out empirically. A study published in the peer-reviewed All Ireland Journal of Higher Education had students use generative AI to complete reformatted past university assessments, aiming to pass when graded by an academic while evading an AI-writing detector; 23.8% of the AI-generated solutions passed (grades of 40-60%) while going undetected. Separately, RTÉ reported roughly 500 instances of AI-enabled cheating detected across Irish higher education in the 2024/25 year, with wide variation between institutions that do and don't track AI-specific cases separately: University of Galway 224 cases, DCU 223 (2025), TU Dublin 71, National College of Ireland 68, University College Cork 22 of 46 total misconduct cases, and SETU 12.

The Higher Education Authority still does not require institutions to record AI misuse separately from other misconduct, so the true national scale likely exceeds what is reported. This is a genuinely cautionary case: it documents the limits, not the success, of AI-based integrity enforcement, and models a transparent, evidence-led policy response instead of overclaiming what the technology can do.