Emotional Dynamics in Digital Learning during the Covid-19 Era: Exploring Generative AI Technologies in Social Research

The widespread adoption of artificial intelligence (AI) tools offers new analytical capabilities that enhance our understanding of social phenomena (Airoldi, 2022). In particular, certain AI models facilitate the identification of trends and patterns in emotions, which are valuable for investigating the mechanisms underlying social structures and group dynamics. The significant integration of digital tools in pedagogical practices, accelerated by the Covid-19 pandemic (Fasanella A., Lo Presti V., Parziale F., 2020), presents both challenges and opportunities for research on emotions and learning within digital environments. Indeed, recent studies in neuroscience demonstrate that human cognitive processes, including attention (Vuilleumier, 2005), learning and memory (Phelps, 2004; Um et al., 2012), and reasoning (Jung et al., 2014), are shaped by emotions. Consequently, it is essential to consider the capacity of educational technologies to influence emotional experiences, with potential implications for learning outcomes (Chai M. Tyng et al., 2017). Based on data collected from the research project "The Evaluation of the Social Impact of Distance Learning after Covid-19" conducted by Sapienza University of Rome, this study aims to analyze the emotional responses of teachers and students in digital learning spaces during the Covid-19 pandemic. This analysis will be conducted using the textual content of interviews and focus groups, transcribed in Italian, and examined through textual analysis techniques employing generative AI. Specifically, the umBERTo AI model, which is derived from the BERT architecture, will be used to conduct sentiment and emotion analysis on Italian-language texts. In light of the findings, this research further seeks to explore the opportunities, limitations and ethical considerations inherent in the application of AI technologies in social research.