Artificial Intelligence In Task-Based Education

Abstract

This thesis examined how artificial intelligence can reduce the workload of educators and students in task-based learning. The study identifies the main challenges of TBL through a literature review. Academic articles and official reports were analysed to identify challenges for educators and students in higher education, with a focus on task-based learning. Considering these challenges, an AI-driven learning platform, Eduhens, was designed, developed and tested. Eduhens was designed using an iterative user-centred process. Starting from a user flow diagram, low- and high-fidelity prototypes were crafted. The platform was designed with the core features of AI-powered task creation, submission tracking and AI-powered assessment. The final design was developed using a modular approach that combined Next.js, Express, Supabase, and GPT-4o-mini. The platform was tested by mimicking educators and students. The results showed that AI can reduce the workload of TBL by accelerating task creation, tracking, and assessment. The tests revealed system problems, such as providing inaccessible resources, generating incompatible rubrics, and issuing evaluations even when the submission link is not accessed. In addition, the limitations of AI, such as cost, credibility, and data safety, were addressed. Despite these constraints, the study concludes that AI can reduce the workload for TBL. With further improvements, Eduhens has the potential to become a dependable support tool for task-based learning.