Review of research on applications of speech recognition technology to assist language learning

Abstract

Abstract Speech recognition technology (SRT) is now widely used in education because of its potential to aid learning, particularly language learning. Nevertheless, SRT has received only limited attention in earlier review studies. The present research aimed to address this gap in the field. To this end, 26 articles published in SSCI journals between 2014 and 2020 were selected and reviewed with respect to domain and skills, technology and their application, participants and duration, measures, reported results, and advantages and disadvantages of SRT. The results showed that English received much more attention than any other language, and scholars mostly focused on facilitating pronunciation skills. Dragon Naturally Speaking and Google speech recognition were the most popular technologies, and their most frequent application was providing feedback. According to the results, college students were involved in research more than any other group, most studies were carried out for less than one month, and most scholars administered a questionnaire or pre-/posttest to collect the data. Positive results related to gains in proficiency and student perceptions of SRT were identified. The study revealed that improved affective factors and enhanced language skills were advantages, whereas a low accuracy rate and insufficiency (i.e. lack of some useful features to support learning efficiently) of SRT were disadvantages. Based on the results, the study puts forward several implications and suggestions for educators and researchers in the field.