Stakeholders’ discourse on generative AI in higher education: Insights from topic modeling and sentiment analysis

Owing to their remarkable proficiency in academic tasks, generative AI (GenAI) systems have been widely embraced in the education sector, stimulating debate among academics about their opportunities and challenges. However, extant studies tend to be narrowly focused, typically drawing on micro-level observations from small samples of students and educators, which limits the scope of reported findings. The present study contributes to the literature by capturing stakeholders’ discussions on the role of GenAI in higher education at scale. It does so by employing over 230K tweets shared on the subject of GenAI and education and training a topic modeling algorithm on this corpus. The model identified seven major themes that encapsulate the use cases and implications of GenAI adoption in higher education as articulated by stakeholders: learning assistant, research tool, productivity tool, assessment and examination, educational technology, adoption policy, and resources. Additionally, the study performed sentiment analysis to reveal the specific themes on which stakeholders expressed disproportionately positive and disproportionately negative attitudes toward GenAI adoption. By combining stakeholder theory and a big data approach, the current findings highlight both the practical and institutional concerns of GenAI beyond immediate classroom implications, such as student learning and assessment.