Development, validation, and refinement of the GenAI competence scale for informal digital language learning: a network and machine learning approach

The entrance of generative artificial intelligence (GenAI) into informal digital language learning brings a new set of linguistic and strategic demands. Learners have to develop new forms of competence in interpreting, evaluating, and making use of machine-generated output for their own independent learning. This suggests the necessity for a valid and context-sensitive framework and instrument of GenAI competence that can support pedagogical actions to balance their collaboration with GenAI. The present study takes up this initiative by proposing a domain-specific framework of GenAI competence and building and validating an instrument for gauging such competence. Study 1 developed and validated the 24-item scale via exploratory factor analysis, exploratory graph analysis, and confirmatory factor analysis. The GenAI competence draws on four interrelated dimensions: knowledge, skills, attitudes, and management. Study 2 generated a short version with 12 items by using genetic algorithms. This study contributes to the diagnosis of learner GenAI competence and the development of instructional practices that promote autonomous learning in the epoch of GenAI.