An article co-authored by a lecturer of the Department of General Pedagogy has been published in Trends in Higher Education.
The study tested whether a digital tool can detect university teachers’ emotional states and respond with supportive messages grounded in Self-Determination Theory.
A few-shot large-language-model classifier performed best at recognising emotions, reaching 0.893 accuracy on a set of faculty-oriented scenarios. Thirty faculty members and researchers rated the messages blind.
Messages that used the full situational context showed the clearest gains in relevance and motivational usefulness. The authors stress that the results show technical and methodological feasibility rather than real-world effectiveness.
Authors: J. Akhmetova (Department of General Pedagogy), O. Kuznetsov, N. Oshanova, A. Zhilkishbayev
