Assessing Speaking Skills Through E-Assessment: A Systematic Review for Advancing Language Evaluation in Indonesian Language Education

Abstract

Speaking skill assessment is a vital component of Indonesian language learning; however, its implementation often faces persistent challenges, particularly in capturing the nuances of students’ spoken expressions and vocabulary use. This study proposes an innovative approach by integrating Artificial Intelligence (AI) as a supportive tool for speaking assessment. Drawing on speech recognition technology and natural language processing (NLP), AI enables the automatic recording, transcription, and analysis of learners’ oral performances. Through a literature review of Scopus-indexed sources, this conceptual research examines the application of e-assessment in speaking skills, identifies the technologies and platforms involved, and evaluates both advantages and challenges. Findings suggest that AI-assisted assessment offers significant benefits, including improved objectivity, reduced teacher workload, enhanced time efficiency, and the provision of timely, in-depth feedback. Nonetheless, limitations remain, such as the need for dialect-sensitive models, infrastructure readiness, and adherence to ethical data practices. This study concludes with recommendations for future development, emphasizing contextually relevant AI systems, teacher training, and ethical frameworks to ensure fair, accurate, and culturally responsive speaking assessment in the Indonesian language context.