AI-Driven Teaching: Exploring The Potential Of AI(LLM)-Based Evaluation For Moodle In The Indian EdTech Market

Abstract

The adoption of Artificial Intelligence (AI) in education is transforming digital learning environments, with Learning Management Systems (LMS) such as Moodle emerging as key platforms for AI-driven teaching and assessment. This study examined the integration of Large Language Models (LLMs) in Moodle, with a focus on the perceptions, usability, and effectiveness of educators and students within Indian educational contexts. A survey of 125 respondents, including educators, administrators, students, and Moodle developers, was conducted to collect quantitative data on usage patterns, perceptions of AI effectiveness, grading consistency, and LLM integration needs. Descriptive statistics, ANOVA, correlation, and Chi-square tests were employed to analyse the data. Key findings reveal that respondents perceive AI tools as moderately effective in addressing student evaluation needs (M = 3.49) and LLM-based grading as consistent compared to manual methods (M = 3.57). The importance of local content customisation (M = 3.52) was also highlighted. ANOVA results indicate significant differences in perceptions of AI effectiveness across user groups, while LLM grading consistency was viewed uniformly. Correlation analysis showed a significant positive relationship (r = 0.470, p < 0.01) between personalised LLM features and the need for local content customisation. Chi-square analyses revealed that perceptions of LLM scalability are influenced by manual grading challenges (p = 0.006) but not by Moodle usage experience (p = 0.944). No significant differences were observed regarding the perceived requirements for effective LLM integration across respondent roles or Moodle experience. The research found that LLMs can provide greater consistency in grades, personalised feedback to students, and scalability within the educational system, but adoption requires school infrastructure and teacher training. The implications related to contextual, localised AI solutions that can be integrated within Indian classrooms. The study suggests the need for structured professional learning opportunities for teachers, phased implementation of LLMs into classrooms, and a consideration of the policy context for educators to engage ethically and responsibly with inclusive and useful AI within the educational setting.