A systematic review of research on emotional artificial intelligence in English language education

Abstract

In learning English as a foreign language (EFL), students often experience foreign language anxiety. Artificial intelligence (AI) applications that provide emotional support and/or create emotional impacts on student learning, so-called emotional AI applications, have received increased attention. However, there is a lack of a systematic review of studies on emotional AI in EFL education. This paper presents a systematic review of research in this field. The results reveal five affordances of emotional AI in EFL education, namely (1) enabling human-like conversations, (2) providing personalized real-time feedback or instructions, (3) translating images into English text, (4) generating personalized learning content and tasks, and (5) recognizing and analyzing emotions. The first three affordances are more frequently used and have shown promising effects on improving students’ behavioral, cognitive, and affective learning outcomes. Moreover, the findings reveal that emotional support is often integrated with cognitive support; providing emotional support alone may not be enough to support student learning. Meanwhile, providing cognitive support alone can enhance both affective and cognitive learning outcomes. Finally, attention should be paid to the factors that might influence the adoption and effects of emotional AI in EFL education.