A Review of Human and Automated Scoring in Second Language Assessment: Opportunities and Challenges

A Review of Human and Automated Scoring in Second Language Assessment: Opportunities and Challenges

Advances in technology and artificial intelligence (AI) tools have led to increased efficiency and innovation in second language (L2) assessment. The development of AI-integrated automated scoring systems is a prominent feature that enables rapid, consistent (re)scoring, reduces costs, provides timely reporting, and offers instant, tailored feedback, especially in large-scale assessments (Zhang, 2013). Despite the new opportunities presented by automated scoring, it also introduces challenges, such as difficulties in assessing cognitively demanding language features (e.g., content, creativity). Therefore, it is important to review the theoretical and empirical research in the literature to understand current technological advances and their future directions. This article first reviews the changes and benefits that technology-based assessments have brought to L2 assessment in general. It then focuses on automated scoring and feedback systems, highlighting the strengths and weaknesses of both human and automated scoring. More specifically, it reviews empirical research on L2 writing and speaking assessments and discusses the comparability of the two. The review identifies gaps in the previous research and suggests future directions for the use of automated scoring.