Objective: To explore the design, implementation, and effectiveness of AI-enabled assessment tools in K-12 language learning through systematic review and meta-analysis.
Methods:
- Systematic review of 25 empirical studies from 2012 to 2024
- Meta-analysis of 21 studies with experimental and control groups
- Data collected from six databases (EBSCO, ProQuest, Scopus, Web of Science, ACM Digital Library, CNKI)
Key Findings:
- Most common design: Structural AI architecture with AI language features + AI algorithms
- Formative-iterative tools were most prevalent, especially for writing assessment
- Short-term interventions (1-8 weeks) showed better results than longer durations
- Secondary school students benefited more than primary school students
- Tools were equally effective for both first and second language learners
Implications:
- AI-enabled assessment tools can significantly enhance K-12 language education
- Need to integrate learning theories with AI algorithms
- Teachers' role evolves from leader to facilitator in AI-empowered environments
- Short-term interventions with clear goals may be more effective
Limitations:
- Small sample size (21 papers) for meta-analysis
- Limited to English and Chinese publications
- Focus only on general language learning outcomes
- Limited timeframe (post-2012)
Future Directions:
- Investigate diverse instructional designs for AI tool integration
- Study long-term impacts in various educational contexts
- Research affective gains across different contexts
- Examine AI assessment in listening and speaking skills
Title and Authors: "A systematic review and meta-analysis of AI-enabled assessment in language learning: Design, implementation, and effectiveness" by Angxuan Chen, Yuyue Zhang, Jiyou Jia, Min Liang, Yingying Cha, and Cher Ping Lim
Published on: 2024
Published by: Journal of Computer Assisted Learning