This study investigates the impact of generative artificial intelligence (AI) on the teaching of French as a Foreign Language (FFL), focusing on AI-generated errors, their effects on learning, and their pedagogical implications. It examines teachers’ uses of generative AI, identifies common types of AI-generated errors, and assesses their influence on learners’ autonomy and critical thinking. A mixed-methods design was adopted, combining a survey of 72 FFL teachers from different educational settings in Morocco with twelve semi-structured interviews. Quantitative and qualitative data were analyzed using descriptive statistics and thematic analysis, respectively, with triangulation employed to strengthen the validity of the findings. Results show that teachers mainly use generative AI for lesson preparation, activity design, and language correction. However, participants identified recurring limitations, including oversimplified content, inaccurate information, grammatical errors, discourse inconsistencies, and cultural biases. They also expressed concerns about excessive reliance on automated correction, reduced cognitive engagement in writing, and declining linguistic autonomy. Nevertheless, AI-generated errors can constitute valuable pedagogical resources by promoting critical thinking, metalinguistic awareness, and learners’ ability to evaluate AI-generated content critically. The study concludes that effective integration of generative AI into FFL teaching requires structured pedagogical mediation. Teachers remain essential in validating AIgenerated content, guiding its educational use, and developing learners’ critical engagement with AI. A balanced integration can therefore support pedagogical innovation while preserving cognitive autonomy and language-learning quality.