Integrating information literacy and AI literacy: a multidimensional framework for critical engagement with generative AI in English language teaching

Purpose This study develops and validates an integrated framework explaining how English Language Teaching (ELT) students critically, ethically and reflectively engage with generative artificial intelligence (AI), operationalized through a 25-item instrument (AI-currency, relevancy, authority, accuracy and purpose (CRAAP)). Design/methodology/approach A cross-sectional design was used with 394 undergraduate students from four Indonesian universities, following a four-stage procedure: construct specification, expert content validation, exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) using maximum likelihood estimation. EFA and CFA drew on the same sample, so CFA results are interpreted as preliminary. Findings Results support a five-dimensional structure: AI usage behaviour, perceived benefits, critical evaluation, ethical awareness and reflective awareness. All five dimensions yielded high mean scores (3.72–3.83, Δ = 0.11), with 65–74% of students classified as high on each dimension. Reflective awareness recorded the lowest mean but remained within the high range, indicating synchronized rather than deficit development. AI engagement in ELT is thus structured through an interdependent evaluative–ethical–reflective continuum. Research limitations/implications The cross-sectional design, self-report measures, Indonesia-specific sampling and single-sample EFA and CFA limit causal inference and full validation; longitudinal, cross-cultural or split-sample designs are recommended. Practical implications The validated instrument gives educators and institutions a diagnostic tool for benchmarking and designing interventions that move beyond regulating AI use toward critical, ethical and reflective engagement. Social implications This study fosters responsible digital citizenship: The framework's emphasis on ethical awareness and reflective awareness prepares students to navigate AI-generated content responsibly. Socially, this translates into graduates who are less likely to engage in plagiarism, misinformation or uncritical acceptance of AI outputs, contributing to a more ethically grounded and discerning online community. In ELT contexts, where language learning shapes identity and communication, the study's finding that reflective awareness is underdeveloped implies a social risk of students passively internalizing biased or inaccurate AI-generated English content. By promoting critical evaluation alongside AI use, the paper socially empowers learners to question AI authority, preserve their own voice and avoid linguistic or cultural homogenization. Originality/value This study offers a novel integrative framework that unifies fragmented competency models, positioning AI literacy as an evolved, evaluative–ethical–reflective form of epistemic literacy validated for ELT.