Task–Based AI Learning Games: A Practical AI Learning Approach

Abstract

Currently, many game-based methods are used in AI education. However, few of them start with requirement analysis and automatically evaluate AI models created by the learner. Therefore, we propose a task-based learning method using LLM (Large Language Model), in which a problem is created in natural language, a solution is described, and coding is performed. We also propose a game, LLM-PoweredAILearningQuest' (LLM-ALQ), which enables this task-based learning. The effectiveness of learning this game and task-based learning was measured. As a result, a certain learning effect was expected for the task-based learning method. On the other hand, issues such as the game's UI and system improvements were also highlighted.