Foster or Hinder Learning? The Role of Generative Artificial Intelligence in the Self-assessment Process

Self-assessment can improve both academic performance (e.g., Brown & Harris, 2013; *Yan et al., 2022, 2023) and self-regulated learning (e.g., Panadero et al., 2017; *Yan et al., under review). However, as students' self-assessments often lack timely feedback (e.g., from a teacher), they can be inaccurate and fail to provide references for critical judgement, which reduces their potential to enhance long-term learning capacities (*Yan & Carless, 2022). As many teachers lack the time to provide such feedback, Generative Artificial Intelligence (GenAI) tools could possibly give such immediate feedback, enabling students to self-assess anytime, anywhere. This gives GenAI the potential to redefine feedback sources and reshape the way students interact with feedback in self-assessment. However, if GenAI gives incorrect, unclear, or confusing information, or if students use it inappropriately (e.g., to avoid making their own judgements), it can hinder their learning. This project explores GenAI's role in the self-assessment process in higher education. By examining the interactions between students and GenAI during the self-assessment process, we will address four specific research objectives (ROs): RO1. Develop a framework for evaluating the quality of GenAI-assisted self-assessments; RO2. Identify ways in which students use GenAI for self-assessment; RO3. Investigate the relationship between students' use of GenAI for self-assessment and the quality of their self-assessments; and RO4. Examine the impact of prompt training on the quality of GenAI-assisted self-assessment.

Funding Source: RGC - General Research Fund (GRF)