Article: Transforming English language education with AI-driven deep learning models for scalable adaptive and inclusive assessment Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.12 pp.15 - 31 Abstract: This study explores how deep learning can help English language education overcome the disadvantages of conventional means of assessment that are typically slow, subjective and not easily scalable. The research uses modern NLP models such as BERT and GPT to create AI-powered systems that assess reading, writing, listening, and speaking skills. They offer personalised real-time feedback designed specifically for any learner. A mixed methods approach combines educator and student insights with performance metrics. Results confirm that AI nearly doubles the accuracy, halves the grading time, and increases note engagement by 50% with clearly gained student proficiency. The study deals with the issue of cultural bias and privacy but uses it ethically and inclusively. By showing deep learning's great promise for providing fair, scalable and effective language learning, this multimodal framework represents an archetype they say can help overcome that challenge. Inderscience Publishers - linking academia, business and industry through research

Title: Transforming English language education with AI-driven deep learning models for scalable adaptive and inclusive assessment

Authors: Yanwen Chen

Addresses: Shanxi Institute of Technology, Yangquan City, Shanxi Province, 045000, China

Abstract: This study explores how deep learning can help English language education overcome the disadvantages of conventional means of assessment that are typically slow, subjective and not easily scalable. The research uses modern NLP models such as BERT and GPT to create AI-powered systems that assess reading, writing, listening, and speaking skills. They offer personalised real-time feedback designed specifically for any learner. A mixed methods approach combines educator and student insights with performance metrics. Results confirm that AI nearly doubles the accuracy, halves the grading time, and increases note engagement by 50% with clearly gained student proficiency. The study deals with the issue of cultural bias and privacy but uses it ethically and inclusively. By showing deep learning's great promise for providing fair, scalable and effective language learning, this multimodal framework represents an archetype they say can help overcome that challenge.

Keywords: deep learning models; natural language processing; NLP; English language education; AI-driven assessments; personalised feedback; multimodal language evaluation.

DOI: 10.1504/IJICT.2025.146163

International Journal of Information and Communication Technology, 2025 Vol.26 No.12, pp.15 - 31

Received: 19 Feb 2025
Accepted: 01 Mar 2025
Published online: 08 May 2025 *

Title: Transforming English language education with AI-driven deep learning models for scalable adaptive and inclusive assessment

Keep up-to-date