Privacy-Preserving Intelligent Tutoring System Using Local AI Models for Personalized Online Learning

Abstract

The rapid evolution of digital technologies has significantly transformed the education sector, leading to the emergence of online learning platforms. However, most existing platforms lack personalized learning experiences and raise concerns regarding data privacy due to cloud-based processing. This research proposes a privacy-preserving intelligent tutoring system that utilizes local artificial intelligence models to deliver personalized educational content without compromising user data.The proposed system is designed as an interactive AI-powered learning platform that operates entirely on the user's local machine. It integrates large language models through local inference frameworks, enabling real-time interaction without the need for external servers. This ensures complete data privacy, as no user information is transmitted outside the system.The platform offers two primary modes: topic explanation and quiz generation. In explanation mode, the system acts as a tutor, providing structured and step-by-step explanations tailored to the user’s educational level and subject. In quiz mode, the system generates customized multiple-choice questions with answers and explanations, enabling users to assess their understanding.The system is implemented using Python and Streamlit for the graphical user interface, providing a seamless and interactive experience. It integrates with local AI models through an inference engine, allowing dynamic selection of models based on availability. The system intelligently prioritizes models for optimal performance and provides feedback to users regarding model selection.One of the key features of the proposed system is its adaptability. Users can select their education level, subject, and learning mode, allowing the system to tailor responses accordingly. This personalized approach enhances learning outcomes by addressing individual needs and preferences.The system also incorporates session management to maintain conversation history, enabling contextual understanding and continuity in interactions. Streaming responses from the AI model ensure real-time feedback, improving user engagement. Experimental observations indicate that the system effectively delivers personalized learning experiences while maintaining high levels of privacy and efficiency. The use of local models eliminates dependency on internet connectivity and reduces latency, making the system suitable for offline learning environments.This research contributes to the field of educational technology by providing a practical solution for privacy-preserving AI-based tutoring. The proposed system addresses key challenges in online learning, including personalization, accessibility, and data security. Future work can focus on integrating advanced adaptive learning algorithms and expanding subject coverage to further enhance the system’s capabilities.