<![CDATA[Math anxiety is an emotional factor that significantly affects students' learning processes and outcomes, particularly in junior high school. Children with this disorder may have difficulty concentrating, have low self-esteem, and are more likely to be passive learners. Therefore, learning innovations are needed that can encourage a more stimulating and enjoyable learning environment. The purpose of this study was to collect the findings of previous research on the integrative potential of ChatGPT-assisted Discovery Learning models in lowering mathematics-related junior high school students' anxiety. The PRISMA protocol was followed in the methodology of the Systematic Literature Review (SLR) of this study. Data for scientific publications published between 2019 and 2025 were collected from Google Scholar and Garuda databases. Regarding the year range, the articles studied focused on a combination of three variables, where articles in the period before the launch of ChatGPT (2019–2022) were netted based on the study of AI chatbots or similar artificial intelligence technologies in mathematics education, while the articles for the period 2023–2025 focused directly on the use of ChatGPT. A total of 15 articles that met the inclusion criteria were reviewed qualitatively based on the focus of mathematics anxiety, the use of Discovery Learning, and the use of AI/ChatGPT. The findings suggest that internal factors (fear and emotional stress) as well as environmental factors (monotonous learning) contribute to math anxiety. The application of Discovery Learning can increase students' freedom and activities, while ChatGPT offers potential support in the form of brief explanations, quick feedback, and personalized assistance. The findings show that the combination of Discovery Learning and ChatGPT has strong integrative potential as an alternative strategy to mitigate the math anxiety of junior high school students through the creation of interactive, independent, and adaptive learning spaces.]]>
