Topic modelling yielded 16 topics dominated by four: AI in Education & Pedagogy (39.43%), Applied ML & Data Science (21.33%), AI Ethics Risk & Governance (20.83%), and ChatGPT for Teaching, Writing & Coding (7.13%)¶

Claim · All claims
Evidence · 1 study · associational r1 · q2
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? After outlier reassignment and topic reduction, the top four topics covered most of the corpus, with Topic 1 at 39.43%, Topic 2 at 21.33%, Topic 3 at 20.83%, and Topic 4 at 7.12%. → Bagale 2025

Evidence¶

Bagale 2025¶

Bagale, A.; Mehjabin, N.; Ünlü, A.; Johri, A. (2025). A Longitudinal Analysis of Public Discourse on AI Ethics in Education Using Twitter Data. https://arxiv.org/abs/2607.12295 doi:10.48550/arxiv.2607.12295

q2 · i? · associational · r1

BERTopic analysis of the tweet corpus after c-TF-IDF outlier reassignment (outliers reduced to 0.08%) and reduction to 16 topics. The article reports "Topic 1 was 39.43%, Topic 2 was 21.33%, Topic 3 was 20.83%, and Topic 4 was 7.12%."

"Topic 1 was 39.43%, Topic 2 was 21.33%, Topic 3 was 20.83%, and Topic 4 was 7.12%. All other topics were below 3% of the total number of tweets."

Discussion¶

Related Claims¶

• AI ethics research is distributed across applied disciplines, with Social Sciences the largest subject area (52.84%) ahead of Computer Science (34.04%) — related
• Discourse on AI in education peaked sharply in 2023 with the rise of generative AI tools such as ChatGPT before declining in 2024 — related
• Algorithmic curation on Twitter may amplify optimistic, technically sophisticated AI-in-education narratives while underrepresenting practical implementation concerns and non-technical stakeholders — related
• ChatGPT research in programming education prioritizes classroom practice and learner interaction, with limited attention to assessment design and governance — related
• Text mining of 69 Scopus documents on ChatGPT in programming education identifies four dominant research themes — related
• University AI policy themes fall into three main categories: usage and education, technology and tools, and governance and potential risks — related
• LDA topic modeling of HI/HIM AI policy documents identifies four dominant themes: academic integrity, generative AI use, AI in research, and ChatGPT engagement — related
• LDA topic modeling of 242 LDT syntheses reveals eight thematic areas, with computational tools most prevalent (21.7%) and generative AI integration emerging (8.2%)
• AI/ML fundamentals, agentic development, prompting, and testing are the most commonly documented topics in AI-assisted software engineering courses — related
• Public sentiment toward AI in education on Twitter was predominantly positive (81.65% of tweets) across 2019–2024 — related
• Educator open-ended responses yield four themes: ChatGPT dominance, need for methodological guidance, ethical concerns, and infrastructure barriers — related
• The most reported risks of AI tools in science and chemistry education are ethical issues, including gender and racial bias, hallucinations, copyright infringement, and plagiarism — related
• Equitable-access and digital-inclusion themes are minimally represented in HI/HIM AI policies — related
• Topic modeling of open-ended responses revealed five themes centered on ChatGPT's simplicity, speed, and role as a support tool — related
• BERTopic analysis of 1,048 AIOL publications identifies research themes including adaptive learning systems, sentiment analysis, and predictive analytics — related
• GenAI guidelines at top U.S. universities cluster into four topics: learning and assessment integration, visual/multimodal media, security and ethics, and academic integrity — related
• GenAI education discourse on Reddit spans 17 themes across 5 clusters, with Academic Integrity the largest at 37.1% of records — related
• Impression narratives rely on weakly diagnostic suspicion heuristics such as response speed, fluency, and perceived scriptedness — related
• Research on SECs and engagement is unevenly distributed across the five CASEL domains, concentrating on self-management, self-awareness and relationship skills — related
• Topic modeling reveals 20 topics in six communities, with special education and collaborative knowledge construction thematically isolated — related