Higher education in Indonesia is facing the challenge of producing graduates who are not only academically competent but also equipped with practical skills relevant to the demands of Industry 4.0. To address this, student-centered learning has been promoted to enhance 21st-century skills such as critical thinking and problem-solving. In this context, Generative AI presents a promising tool to support learning and research activities, helping students design survey instruments, analyze data, and structure research reports more efficiently. Despite growing interest in the technical and pedagogical aspects of Generative AI, few studies have examined its impact on students’ quantitative research skills and research self-efficacy. This study integrates Generative AI into research-based learning within an Applied Statistics course, implementing the 5E Learning Model (Engage, Explore, Explain, Elaborate, Evaluate) with ChatGPT as a learning aid. Using the DeLone & McLean model, Technology Acceptance Model, and Rational Choice Theory, this research explores students’ perceptions, adoption drivers, and the perceived impact of Generative AI on research performance. Data collected from 219 undergraduate students indicate that perceived usefulness, ease of use, and satisfaction significantly influence students’ intention to adopt Generative AI, while perceived risk and learning cost do not. Moreover, Generative AI usage contributes positively to students’ research performance, including improved output quality, enhanced research skills, greater self-confidence, and a stronger intention to undertake a research project for graduation. These findings not only enrich the literature on educational technology adoption but also offer practical insights for integrating Generative AI into higher education curricula to support digital transformation in learning.