The rapid evolution of machine translation and generative AI in recent years—driven by big data—has resulted in output that closely approximates natural human language. Against this backdrop, a growing body of literature has advocated for the integration of such technologies into English language education. Nevertheless, practical classroom-based applications remain limited, and few studies have provided concrete evidence to substantiate the pedagogical benefits of machine translation. This study aims to evaluate the effectiveness of machine translation as a tool for learning English writing by examining cerebral hemodynamic responses and learners’ actual writing performance. Japanese university students with TOEIC scores in the low 600s were divided into two groups: those who used machine translation and those who did not. The analysis revealed that activating the brain region associated with language production—specifically, Broca’s area—may be essential for fostering writing proficiency. Although the use of machine translation was found to have limited influence on the activation of this area, it showed potential utility in facilitating vocabulary acquisition. These results suggest that while machine translation may not fully support the cognitive processes underlying syntactic production, it can serve as a complementary tool promoting language learning.
