To what extent can the critical analysis of errors produced by artificial intelligence foster the development of metalinguistic competence among learners of French as a foreign language?

Abstract

This study examines the potential of artificial intelligence-generated errors for developing metalinguistic competence in French as a Foreign Language (FLE). It is based on an experiment conducted with 30 students over six weeks, during which learners were asked to identify, correct, and justify grammatical, lexical, and syntactic errors produced by AI. A comparison between a pre-test and a post-test was used to assess changes in their performance. The results show improvement in error detection, correction, and, particularly, the ability to justify linguistic choices. The study highlights the value of using AI as an object of analysis rather than merely as a correction tool, thereby encouraging a more critical and reflective approach to language learning.