Beyond the Desktop: Assessing Computer-Based Reading Comprehension in the Age of Ubiquitous AI

Abstract

Language education in the digital world has moved beyond simple desktop learning tasks to a rich ecosystem of pervasive Artificial Intelligence (AI) and Mobile Assisted Language Learning (MALL). In this research, the researcher investigates the computer-based reading comprehension in this transformative era with emphasis on Intelligent Computer-Assisted Language Learning (ICALL). The research is based on pedagogical theories such as Merrill's First Principles of Instruction (FPI) and Activity Theory, which guide the examination of how AI agents act as socio-pedagogical scaffolds. Results have shown that AI can boost personalization by adapting content, making it easier or more challenging based on the individual's level of proficiency. They also help in incidental vocabulary and enhance learner engagement. The use of generative models such as ChatGPT can lower the language learning anxiety and increase the level of learner autonomy. But even with these machines, there are still “data failures” when they work to "understand" the subtle socio-cultural context. Also, there are some technical problems such as small screen size, which could make MALL less effective than CALL in the intensive reading activities. The research results indicate that in order to evaluate comprehension, holistic interaction between the learner and the intelligent agent must be analyzed. Implementing this successfully requires the ability to compensate for the computational ability of AI with the qualitative understanding of human educators.