A Dialogue Between Students’ Imagination and GenAI’s Visualisation of the Text: Using AI Image Generator to Foster Critical Thinking in the Reading Process

Abstract The study explores the use of Generative Artificial Intelligence (GenAI) for visualisation to foster students’ critical thinking and reading comprehension in second language (L2) reading. Using a task based approach in which students needed to read a text, identify the key ideas, and make prompts asking GenAI to create images to reflect the main contents of the text, the study then asked learners to reflect on the task and images created by AI. There were 32 undergraduate students whose English level was from B2 to C2 (CEFR) participating in the research. Applying surveys to collect students’ opinions, the study analysed their ideas using qualitative content analysis with the application of Excel, adopting phenomenology as the philosophical stance. The findings indicate that approximately one third of the students perceived GenAI-generated images as lacking accuracy and offering limited variation despite different prompts. They also noted that the images tended to appear idealised, even when the source text conveyed negative realities. In contrast, about one fifth of the participants found the images helpful in making the text more concrete and easier to understand, while a few acknowledged that they expanded imaginative possibilities. However, expecting GenAI to replicate human mental imagery may generate extraneous cognitive load. These findings suggest that GenAI should be used not as a tool for exact reproduction of human imagination, but as a means to extend knowledge, foster critical engagement with both text and image, and encourage evaluative feedback that supports human-AI co-creation of meaningful outcomes.