Leaving Cert Exam Reforms Delayed Due to Concerns of AI Cheating, Ireland
Plans to have teachers assess students for 40% of marks in Leaving Cert exams have been put on hold due to fears of cheating facilitated by artificial intelligence (AI) platforms like ChatGPT. Education Minister Norma Foley has announced an accelerated rollout of senior cycle changes, including the introduction of nine new or revised subjects. However, the proposal for teacher grading has been delayed to allow the State Examinations Commission (SEC) to examine the potential impact of AI on teacher-based assessments.
The revised schedule for Leaving Cert reforms means that students entering fifth year in 2025 will have the opportunity to engage with new subjects and specifications two years earlier than originally planned. The first assessments in these subjects will take place in 2027 and will include additional components, such as projects, in addition to the traditional June exams. The aim of the reforms is to alleviate pressure on students by introducing various forms of assessment over an extended period, as opposed to relying solely on a single written exam in June.
However, until further research on generative AI (GenAI) is conducted, the SEC will continue appointing examiners to assess these additional components, following existing practices for orals, practicals, and projects. This decision comes as a result of concerns about the potential threat to academic integrity posed by AI platforms like ChatGPT, which can rapidly generate convincing and plausible text, images, and other outputs based on prompts.
Education Minister Norma Foley acknowledged the need to explore the opportunities and challenges presented by the advancements in generative AI. She emphasized that the National Council for Curriculum and Assessment (NCCA) would develop externally assessed components for each subject that go beyond traditional written exams. While the SEC already appoints examiners for many Leaving Cert subjects, the minister originally intended for class teachers to take on this responsibility. However, due to the AI-induced risks to integrity, the proposal has been postponed pending further research.
The rollout of GenAI platforms, such as ChatGPT, less than a year ago caused a significant impact in the education sector as they can generate text, images, and code that appear authentic and plausible. While GenAI has potential as an educational tool, inappropriate use can undermine student learning and academic integrity. Quality and Qualifications Ireland (QQI), the academic standards watchdog for higher and further education, recently issued guidelines for educators and students about GenAI. QQI has advised institutions to review assessment approaches to ensure they cannot be completed satisfactorily by individuals using AI without a sufficient understanding of the subject, or that are vulnerable to breaches of integrity or security.
The guidelines recommend avoiding unsupervised online testing and at-home essays or assignments that primarily focus on subject knowledge. Instead, multiple forms of assessment, including greater use of oral exams, are suggested to evaluate integrity and student understanding. While the guidelines discourage a rushed return to supervised written exams, they promote a balanced approach that incorporates various assessment methods.
The decision to postpone teacher grading in Leaving Cert exams is a response to the potential risks associated with AI-enabled cheating. By prioritizing research on generative AI, the Education Minister aims to ensure the integrity of the assessment process while providing opportunities for students to engage with new subjects and specifications earlier than planned. As AI continues to evolve, it is crucial to strike a balance between leveraging its educational potential and safeguarding academic integrity.