Rethinking Language Education After Crisis: AI-Driven Learning Analytics as a Site of Social Science Inquiry European Scientific Journal ESJ

https://www.academia.edu/171637924/Rethinking_Language_Education_After_Crisis_AI_Driven_Learning_Analytics_as_a_Site_of_Social_Science_Inquiry Promising to give personalised feedback and monitor learner behaviour, artificial intelligence-driven learning analytics (AI-LA) gained prominence in language education during the COVID-19 pandemic. However, its adoption has grown faster than the social science research needed to make sense. Conceptually, a structured integrative synthesis of recent peer-reviewed literature is conducted following Jaakkola’s (2020) theory-synthesis approach, prioritizing work published since 2020, and the regulatory shift marked by the European Union’s 2024 Artificial Intelligence Act. Bringing together critical data studies (Boyd & Crawford, 2012; Williamson, Komljenovic & Gulson, 2024), recent systematic reviews of generative AI and personalisation in language learning (Jeon, 2025; Teng, 2025), scholarship on algorithmic bias and linguistic legitimacy (Baker & Hawn, 2022; Koenecke et al., 2020), and post-crisis education research (Charitonos et al., 2025; Menashy & Zakharia, 2022), it argues that AI-LA should be understood as a sociomaterial arrangement rather than a technical pipeline. Its contribution is a conceptual framework, with testable propositions, showing where the pedagogical promise of personalisation breaks down epistemically, why bias in language-recognition systems is political as much as technical and bound up with whose language counts as legitimate, and how the displaced learner exposes the limits of current AI-LA. Implications for researchers, language teachers, curriculum developers, learners, and policy makers are listed thereafter.