Generative AI for Personalized Learning and Education
Meghana Puvvadi, Sai Kumar Arava, Adarsh Santoria, Sesha Sai Prasanna Chennupati, Harsha Vardhan Puvvadi
Abstract
Personalized learning has significantly enhanced teaching practice through its development to cater to varied learning needs and in its full utilization of learner engagement. Traditional practice, in contrast, is limited by fixed content, minimum learner flexibility, and minimal immediate feedback. Generative Artificial Intelligence (AI) introduces a revolution through self-generated adaptive learning material such as customized tests, interactive simulations, and adaptive lesson plans. This article presents a systematic review on the use of Generative AI in education, synthesizing what has been published on its contributions to learning achievement based on empirical studies. Studies conducted by AI-powered platforms like Khan Academy's AI tutor and Duolingo's adaptive questions indicate a 26-38% boost in student engagement, a 34% boost in learning retention, and a 40% decrease in teacher workloads. Additionally, current evidence shows that adaptive learning models implemented using AI cut dropout rates by 22% and improved evaluation accuracy through live automated feedback. Despite these advantages, data confidentiality, algorithm bias, and material reliability concerns continue to exist. This paper fills this gap through the discussion of ethical AI platforms, the usage of open-sourced datasets in minimizing bias, and Explainable AI (XAI) as a means for transparency. Moreover, it explores hybrid AI-human models of education, scalability of AI-powered education in developing regions, and policy recommendations for affordable education. By integrating existing research, this paper also contributes to the body of knowledge on AI-facilitated personalized learning and its implications for learning in the future.