“AI-Powered CEFR-J Resource Development”

As the Common European Framework of Reference for Languages (CEFR) grows increasingly influential worldwide, there is mounting interest in adopting or adapting it to various contexts. The CEFR-J stands as one of the earliest adaptation projects and has gained recognition for its unique research findings and resources. In this talk, I will briefly introduce the CEFR-J project from a historical perspective, demonstrating how the CEFR was examined and explored as a common framework for language education.

More recently, the advent of AI appears to be transforming the entire language learning paradigm. AI is now robust enough to handle various language-related tasks, potentially becoming an extremely powerful tool to enhance language skills. At the same time, both the potential and limitations of AI must be investigated to determine what AI can do for us and how much we should allow it to do. I would argue that to properly assess AI's role in language learning and teaching, it's crucial to clearly define levels of language proficiency and language tasks involved. This clarity enables us to better evaluate AI capabilities in relation to specific tasks at given proficiency levels.

I will then discuss the recent progress of the CEFR-J project, focusing on how AI enhances the development of various CEFR-J-based tools—including text and grammar profiling, speech-to-text conversion, functional expression analysis, and error correction. These AI-powered tools and resources will help clearly define the areas where AI can provide support under specific contexts and conditions.