Applying AI in the Classroom with Leonardo AI

The pedagogical potential of Leonardo AI in the EFL classroom is significant. First, it democratizes visual production: students who lack drawing skills can still visualise complex scenarios (for example, an urban plan for a sustainable Dream City), focusing their cognitive energy on language use rather than technical execution. Second, image generation naturally invites rich language work — descriptive lexis, comparative structures, and evaluative discourse — as learners negotiate prompt wording and interpret AI outputs. Third, the tool supports AI literacy: iterative prompt-editing provides a concrete context for students to observe how language choices affect machine outputs and to practise critical appraisal of automated artefacts. As EdTechFocus (n.d.) suggests, such activities help learners develop the digital and cognitive competencies needed to interact constructively with automated systems.

From the standpoint of the SAMR-AI model, Leonardo AI often moves activities beyond simple substitution. At the Substitution level it replaces manual drawing tools; at Augmentation it provides immediate visual feedback that supports revision; at Modification it changes task design by enabling rapid multimodal prototyping and peer review; and at Redefinition it enables projects that were previously impractical—such as student-designed AI exhibitions or community campaigns combining student texts and AI imagery (Gillespie, 2022; EdTechFocus, n.d.). Crucially, teachers should plan tasks that foreground pedagogical aims (language practice, critical reflection) rather than novelty effects.

Despite its affordances, Leonardo AI raises important pedagogical and ethical concerns that teachers must address. Generative models may reproduce cultural biases, generate misleading or stereotyped imagery, or fail to represent marginalized perspectives accurately; therefore, outputs should always be subject to critical interrogation (UNESCO, 2023). Overreliance on AI for creative production can also diminish learners’ agency if the tool is used as a passive generator rather than as a collaborative partner; teachers should structure tasks that require students to edit, justify and contextualize AI artefacts. Finally, privacy and copyright considerations must be observed when students upload proprietary text or when generated images are shared publicly.