Research Specifications

The rapid integration of Artificial Intelligence (AI) into educational contexts has opened new pedagogical avenues for enhancing second language acquisition, particularly in the development of productive skills such as oral communication. This study investigates the impact of AI-assisted learning on the speaking performance of Iranian English as a Foreign Language (EFL) learners, alongside a rigorous examination of self-regulation as a potential moderating variable. A quasi-experimental, pre-test/post-test control group design was employed, involving a homogenous sample of 60 intermediate EFL learners. Participants were randomly assigned to either an experimental group (n = 30) that utilized an emotionally intelligent AI conversational assistant, Pi, for structured oral practice, or a control group (n = 30) that received conventional, teacher-cantered speaking instruction over an intensive 12-week period. Data were collected using a validated Self-Regulation Questionnaire (SRQ) and an objective AI-administered speaking proficiency assessment tool, SmallTalk2Me. Multivariate Analysis of Covariance (MANCOVA) was utilized to analyse the data, with baseline pre-test scores serving as covariates. The inferential results revealed that learners engaged in AI-assisted instruction demonstrated significantly higher post-test speaking performance (p <.001, partial η² =.650) and enhanced self-regulation skills (p <.001, partial η² =.706) compared to the control group. Contrary to the initial hypothesis, baseline self-regulation skills did not significantly moderate the relationship between AI-assisted learning and speaking performance (p =.957). These findings suggest that the highly structured, adaptive scaffolding provided by the AI tool benefitted learners uniformly, effectively compensating for prior deficits in self-regulatory capacities. The study underscores the transformative potential of conversational AI in mitigating foreign language speaking anxiety, providing indiv