PGCE Students' Perceptions of Generative Artificial Intelligence within Learning Ecosystems at a South African Teacher Education Institution

Abstract

The Fourth Industrial Revolution has brought artificial intelligence (AI) to the forefront of pedagogy in higher education institutions. However, limited empirical literature outlines students' perceptions of AI-powered learning in resource-poor environments in the Global South. Methods This study adopted an interpretive, mixed-methods, single-case design to investigate undergraduate students' perceptions of AI in their learning ecosystems. A purposive sampling approach was used to purposefully select 88 students taking the ICTs in Education class at a South African higher education institution. All the data collected in this study were based on one comprehensive research questionnaire that included both fixed-response Likert items and open-ended questions. The survey instrument elicited students' general perceptions about AI and their experience in a course activity where the task was to prompt AI to develop lesson plans and produce media, as well as create digital assessment activities using products such as posters, games, or videos. Results show that AI emerged as a powerful digital mediational artefact within the integrated conceptual framework that combines the principles of Vygotskian sociocultural theory and the ConTIS Model. 75% of respondents said that AI-based tasks made them feel more personalised and confident in the process of learning in their Zone of Proximal Development. However, there are several obstacles to AI-driven transformation. Socio-material limitations in the region, particularly data poverty and load shedding, constantly hinder the continuity of AI interactions, in addition to teachers' lack of competence in using AI tools (68%) and data privacy issues (61%).