Efektivitas Artificial Intelligence-Assisted Learning terhadap Kemampuan Pemecahan Masalah Matematis Siswa SMA Negeri 1 Denpasar

The development of generative artificial intelligence technology has opened new opportunities in mathematics learning, particularly in training problem-solving skills among secondary school students. This study aims to evaluate the effectiveness of artificial intelligence-assisted learning on the mathematical problem-solving ability of students at SMA Negeri 1 Denpasar compared to conventional learning. The study employed a quantitative approach with a quasi-experimental non-equivalent control group design involving 63 eleventh-grade students. The experimental group (n = 32) received generative AI-assisted learning (ChatGPT) over eight sessions, while the control group (n = 31) received conventional learning. Data were collected through a validated mathematical problem-solving ability test (content validity and Cronbach's alpha reliability = 0.87) administered at the pretest and posttest stages, and subsequently analyzed using the Shapiro-Wilk test, Levene's test, independent t-test, and N-Gain calculation. The results showed that the experimental group experienced a significantly higher increase in mean score (gain = 23.7) compared to the control group (gain = 12.4), with t = 5.84 (p < 0.001) and a moderate-to-high N-Gain category (0.57). These findings indicate that generative AI-assisted learning is effective in improving students' mathematical problem-solving ability through personalized feedback, adaptive problem-solving simulation, and enhanced cognitive engagement during the learning process.