ICTlogy

Abstract:

With the increasing application of generative AI tools, such as ChatGPT, many schools have developed school-based policies to regulate teachers' and students' use of these technologies. However, empirical evidence on the effectiveness of such policies remains limited. This study examines the relationship between school-based ChatGPT policies and students' computer and information literacy (CIL), and whether these policies moderate CIL disparities across students' gender, socioeconomic status (SES), and immigration background. The data come from International Computer and Information Literacy Study (ICILS) 2023, where school-based ChatGPT policies are distinguished for teachers and for students, each with three categories: unrestricted use, restricted use, and not allowed. Employing cross-level interaction models with random slopes, three findings emerge: (1) female, higher-SES, and non-immigrant students exhibit higher levels of CIL after controlling for student- and school-level covariates, including school-based ChatGPT policies; (2) policies permitting students' unrestricted use are negatively associated with CIL, whereas other policies for students and teachers show no significant direct relationship with CIL; (3) moderation is observed. Policies allowing teachers' use, whether unrestricted or restricted, narrow SES-related CIL gaps, while those permitting students’ unrestricted use exacerbate digital gender divide. These results highlight the need for well-designed and institution-supported AI use policies to promote equitable digital competence development in schools and to prevent the widening of existing digital divides.