Artificial intelligence (AI), particularly large language models like ChatGPT, is rapidly transforming educational practices. ChatGPT offers new opportunities for enhancing self-directed learning and personalized instruction, yet empirical studies on its effectiveness remain limited. This study employed a mixed-methods approach, combining a systematic literature review (SLR) with a quasi-experimental design involving 36 high school students. Participants completed pretests and posttests measuring conceptual understanding, and a perception survey was administered to assess student experiences using ChatGPT as a virtual tutor. Quantitative findings revealed a significant improvement in learning outcomes, with posttest scores increasing by 15.5 points on average compared to pretest scores. Survey results showed that 89% of students reported improved comprehension of difficult concepts through interactions with ChatGPT. Students also noted increased motivation for independent learning and appreciated the immediate feedback provided by the AI tool, which helped accelerate task completion. The findings suggest that ChatGPT can serve as an effective supplementary learning tool, especially in supporting self-paced learning. However, critical challenges—such as AI bias, data privacy concerns, and limited technological access in remote areas—must be addressed to ensure equitable adoption. ChatGPT holds strong potential to enhance educational outcomes and promote more inclusive learning environments. Future research should explore culturally adaptive implementations, teacher training, and ethical safeguards to maximize its effectiveness in diverse educational contexts.
Atikah et al. (Tue,) studied this question.