Text-to-image AI e didattica della produzione scritta: un’attività didattica sperimentale

This chapter presents an experimental classroom activity that integrates text-to-image AI into writing instruction at lower secondary school level. Drawing on a research project conducted in Turin in 2024, the author describes a four-phase approach in which students write short narrative passages, then use DALL-E 3 (OpenAI's image generator), mediated by the teacher, to generate images from their texts. The resulting image serves as visual feedback, prompting students to revise and enrich their descriptions. The activity aims to develop critical thinking, lexical fluency, and revision skills, areas identified as consistently neglected in Italian writing pedagogy. While the approach shows pedagogical promise, the author notes that its effectiveness compared to traditional methods remains to be systematically studied, and that teacher competence with AI tools is essential to its success.