Dataset on Epistemic Negotiation in GenAI-Mediated Writing (YouTube API v3 Analysis)

This dataset comprises a computational extraction of audience reflections and educational discourse (N=200) regarding the integration of Generative Artificial Intelligence (GenAI) in higher education. The data was collected utilizing the YouTube Data API v3, focusing on high-engagement pedagogical discussions. The primary objective of this dataset is to provide empirical grounding for the Digital Collaborative Writing Model (DCWM). It maps key linguistic markers associated with 'epistemic drift', 'AI bias', and 'verification patterns'. By categorizing user-generated content, the dataset validates the necessity of structured pedagogical roles—specifically the 'Challenger' and 'Synthesizer'—as psychological and cognitive buffers against the risks of automated reasoning and cognitive unloading in academic writing. This research aligns with the UN 2030 Agenda (SDG 4) and UNESCO’s Digital Competency Framework, emphasizing critical digital participation over passive tool consumption.