Trust, ethical, and learning impact of artificial intelligence feedback in English education

Background: Artificial intelligence-generated feedback (AI-GF) is increasingly used in computer-assisted English as a Foreign Language formative assessment, yet learner engagement with such feedback varies considerably. Previous research has focused largely on technical effectiveness, providing limited insight into how trust, ethical concerns, and learner engagement shape educational value. This gap restricts informed and responsible integration of artificial intelligence feedback in higher-education language learning. Objectives: This study investigates how trust, ethical concerns, clarity, and learner engagement influence the perceived usefulness and future readiness associated with AI-GF in English as a Foreign Language (EFL) learning. Methods: Drawing on technology acceptance and feedback literacy frameworks, a learner centered conceptual model was developed. Survey data were collected from 255 university-level EFL learners enrolled in higher education institutions. The data was analyzed using Partial Least Squares Structural Equation Modelling (PLS-SEM) to examine relationships among trust, ethics, engagement, clarity, perceived usefulness, and future readiness. Learners' prior experience with artificial intelligence (AI) and academic maturity were included to capture variation in engagement patterns. Results and conclusion: The findings indicate that trust is the strongest predictor of learner engagement with AI-GF, while ethical concerns significantly undermine trust. Engagement operates as the principal mechanism linking AI-GF to perceived usefulness and readiness for future learning, whereas clarity alone is necessary but insufficient to support meaningful learning outcomes. These results extend existing models by positioning engagement as a central outcome and integrating trust and ethics as foundational determinants. Pedagogically, the study highlights the importance of ethical transparency, trust-building practices, and structured learner guidance in maximizing the educational impact of AI-GF in computer-assisted language learning contexts.