<![CDATA[The rapid development of artificial intelligence (AI) has created new opportunities to support innovative learning environments in STEM education. However, many science learning processes still rely on conventional instructional methods that do not fully facilitate students’ higher-order thinking skills. This study aims to develop and evaluate SUG-AI, an artificial intelligence–based learning platform designed to support adaptive STEM learning and improve college students’ critical thinking skills in science education. The research employed a quantitative research design integrated with a development approach using the ADDIE model, which includes analysis, design, development, implementation, and evaluation stages. The developed platform provides interactive STEM learning modules, adaptive learning recommendations, and automated feedback based on students’ learning performance. Expert validation was conducted to evaluate the feasibility of the platform, followed by a limited implementation involving college students in science learning activities. Data were collected using validation instruments and critical thinking tests through pre-test and post-test measurements. The results showed that the SUG-AI platform achieved a high level of feasibility based on expert evaluation in terms of content quality, interface design, learning interaction, and system usability. Statistical analysis also revealed a significant improvement in students’ critical thinking skills after using the platform. These findings indicate that AI-based learning platforms have strong potential to enhance interactive STEM learning environments and support the development of higher-order thinking skills in higher education contexts]]>
