Creative thinking in engineering education: project-based English instruction with generative ai tools

Abstract

Purpose. This study investigates the didactic potential of generative artificial intelligence (AI) tools in project-based English for Specific Purposes (ESP) instruction, with the primary goal of fostering creative thinking competencies in undergraduate engineering students. Methods. A mixed-methods design was employed over one academic semester at Vinnytsia National Technical University. Eighty-four fourth-year engineering students were distributed between an experimental group (n = 42), which utilised generative AI platforms (ChatGPT, DALL·E, Canva AI, Tome, Gamma), and a control group (n = 42) engaged in equivalent project-based tasks without AI support. Creative thinking was assessed using adapted verbal tasks the Torrance Tests of Creative Thinking (TTCT), complemented by rubric-based project artefact evaluation, motivation surveys, and semi-structured interviews. Key results. The experimental group demonstrated significantly greater pre- to post-test gains in creative thinking scores (+9.4 points vs. +2.9 points; p < .001), as well as superior performance in lexical variety, structural coherence, and multimodal originality. Student motivation and willingness to communicate in English were markedly higher, with 81% of participants reporting heightened creative self-efficacy. Conclusions. Generative AI tools, when embedded within structured pedagogical frameworks, function as effective cognitive scaffolds rather than mere productivity aids, enhancing both creative output and language proficiency. However, critical digital literacy and explicit ethical instruction are essential preconditions for sustainable integration. The findings offer a replicable model for modernising ESP curricula in technical higher education, with particular relevance to wartime and post-war Ukrainian academic contexts.