A Longitudinal Analysis of Public Discourse on AI Ethics in Education Using Twitter Data

This study provides a longitudinal analysis of public discourse regarding AI ethics in education by examining 14,201 tweets posted between 2019 and 2024. The researchers utilized a combination of BERT-based topic modeling (BERTopic) to identify thematic clusters and a SetFit-based sentiment analysis classifier to track the emotional tone of the conversation over time. By focusing on this five-year window, the authors aimed to capture how the public conversation shifted in response to major technological milestones, most notably the release of ChatGPT.

The analysis reveals that public sentiment toward AI in education has remained consistently optimistic, with approximately 81.6% of the analyzed tweets classified as positive. Rather than reflecting a deeply polarized debate, the discourse is characterized by a pragmatic focus on the potential benefits of AI integration, such as personalized learning and administrative efficiency.

Negative sentiment, while present, is not the dominant tone; instead, it is concentrated in specific, transient spikes that correspond to ethical controversies or anxieties. Over time, the conversation has evolved from general discussions about data privacy and surveillance to more urgent, specific concerns regarding academic integrity and the long-term implications of generative AI on student learning and critical thinking skills. The authors conclude that while the public is largely receptive to AI, there is a growing, consistent demand for institutional accountability and ethical oversight.

Understanding public sentiment is essential for educators, policymakers, and institutional leaders who are tasked with integrating AI into classrooms. This study demonstrates that the public is not inherently anti-AI but is instead navigating a complex set of expectations. By identifying the specific themes that trigger public anxiety, this research provides an empirical foundation for developing transparent and equitable AI policies that address real-world concerns rather than just technical capabilities.