Adaptive Artificial Intelligence for Personalized Pronunciation and Grammar Instruction in Multilingual African Classrooms: A QuasiExperimental Study of Automatic Speech Recognition and Natural Language Processing

Artificial Intelligence (AI) has emerged as a promising approach for delivering personalized language instruction through adaptive learning technologies. However, empirical evidence regarding the effectiveness of integrated AI-assisted pronunciation and grammar instruction in multilingual African classrooms remains limited. This study investigated the effectiveness of an adaptive AI framework integrating Automatic Speech Recognition , Natural Language Processing , adaptive learning algorithms, and peer-supported learning for improving language proficiency among multilingual learners. A quantitative quasi-experimental pre-test– post-test control group design was employed involving 120 learners from multilingual classrooms in Nigeria. Participants were assigned to four instructional conditions: ASR-assisted IJEE IJEE www.ijee.io pronunciation instruction, AI-assisted grammar instruction, combined ASR–NLP instruction, and conventional teacher-led instruction. Data were analyzed using descriptive statistics, Analysis of Covariance , hierarchical multiple regression, and Pearson correlation analysis. The findings revealed significant improvements in pronunciation accuracy, grammatical proficiency, listening comprehension, and overall speaking performance among learners receiving AI- assisted instruction compared with the control group (all p < .05). The combined ASR–NLP intervention consistently produced the largest learning gains across all outcome measures, indicating synergistic effects beyond those achieved by either instructional approach alone. Learner characteristics, particularly first-language background and digital literacy, partially moderated treatment effectiveness. The study concludes that adaptive AI-assisted instruction provides an effective and scalable approach to personalized language learning in multilingual African classrooms. The findings contribute empirical evidence to Computer-Assisted Language Learning research and support the responsible integration of culturally responsive AI technologies into language education in resource-constrained multilingual contexts