Automatic Speech Recognition (ASR) has emerged as a transformative technology in education, yet its application to children’s speech remains underexplored. This paper presents a systematic review of ASR for children, focusing on its effectiveness in educational contexts. Following the guidelines of Kitchenham and Charters (2007), we analyzed academic articles published between 2019 and 2024, sourced from the ACM Digital Library, IEEE Xplore, and Scopus. A total of 16 articles were selected based on predefined inclusion criteria, including relevance to children’s speech and educational applications. The review identifies Deep Neural Networks (DNNs) and adversarial learning models as the most effective approaches for recognizing children’s speech. Key findings highlight the potential of ASR to enhance language learning and development in children, particularly in low-resource contexts. However, challenges such as data scarcity and the need for adaptation to diverse linguistic environments remain significant barriers. This study contributes to the ongoing discussion on innovative educational technologies by providing a comprehensive analysis of current trends and future directions in ASR for children.
Paradeda et al. (Sun,) studied this question.