Power structures and perceptions of AI fairness in high-stakes language testing: the Pearson Test of English (PTE Academic) as case study

Educational assessment is currently undergoing extensive transformation, with the integration of Artificial Intelligence (AI), raising critical questions about fairness and validity. This article applies Ajzen's Theory of Planned Behaviour (TPB) to provide a micro-level analysis of how individual attitudes, subjective norms, and perceived control shape test-takers’ interactions with AI-based assessments, using the Pearson Test of English (PTE Academic) as a case study. Findings suggest that candidates’ views on AI significantly affect their test-taking strategies. Theoretical insights from Foucault are then synthesised to provide a macro-level analysis. This hybrid analysis approach contextualises the socio-political situation of AI in relation to educational testing, showing how it may reinforce existing structures of privilege. Implications for educational practice are developed in the final section.