A systematic research synthesis of generative AI literacy in English language education
TL;DR
Overall, self-report supports perceived-literacy claims, task evidence supports situated-use claims, and broader multidomain claims require aligned constructs, measures, and evidence.
Abstract
Generative artificial intelligence (GenAI) is increasingly embedded in English language writing, feedback, communication, and teacher decision-making, yet GenAI literacy remains unevenly conceptualized and evidenced. This systematic research synthesis reviews 27 empirical studies available between 2024 and March 2026 to examine how GenAI literacy is conceptualized, operationalized, and represented in English language education. Following PRISMA-informed procedures, the review used a provisional three-domain lens covering conceptual understanding, strategic application, and critical–ethical evaluation and conducted a study quality and evidentiary alignment appraisal. Strategic application was the most frequently evidenced domain, and critical–ethical evaluation was also widely visible, whereas conceptual understanding was less often assessed directly. Twenty-two studies provided direct domain evidence; five provided indirect domain-relevant evidence without direct domain coding. Among the 20 studies directly evidencing two or more domains, cross-component coordination was common. Perceived literacy, engagement, and language outcomes remain informative, although they do not automatically demonstrate literacy. Overall, self-report supports perceived-literacy claims, task evidence supports situated-use claims, and broader multidomain claims require aligned constructs, measures, and evidence.
