<![CDATA[The rapid advancement of Artificial Intelligence (AI) has transformed educational practices, particularly in mathematics education, by enabling adaptive learning, personalized instruction, and instant feedback. While AI-powered technologies such as ChatGPT, Gemini, and Intelligent Tutoring Systems have been reported to support students' conceptual understanding and learning experiences, concerns have emerged regarding excessive reliance on AI, which may reduce mathematical reasoning, critical thinking, and learning autonomy. Although prior reviews have examined the general effectiveness of AI in education, few have specifically synthesized how AI reshapes mathematical literacy, and none has systematically addressed this issue within the context of Generative AI. This study addresses that gap by systematically examining the role of AI in the development of mathematical literacy, identifying its benefits and challenges, and exploring future research directions in mathematics education. A Systematic Literature Review (SLR) was conducted following the PRISMA 2020 guidelines. Literature was retrieved from Google Scholar, ERIC, Scopus, and Garuda databases using predefined keyword combinations related to AI and mathematical literacy, and screened against explicit inclusion and exclusion criteria. From 287 identified publications, 32 studies published between 2015 and 2024 met the inclusion criteria and were analyzed using thematic analysis. The findings indicate that AI is reported to contribute to mathematical literacy by supporting personalized learning, adaptive feedback, conceptual understanding, and problem-solving skills, although the evidence is not uniformly consistent across contexts, AI tools, and methodological designs. At the same time, excessive AI use may encourage cognitive offloading, reduce independent reasoning, and increase students' dependence on automated solutions. Beyond synthesizing existing evidence, this review proposes that mathematical literacy in the AI era should extend beyond traditional competencies to include the ability to critically evaluate AI-generated outputs, identify algorithmic limitations, and use AI responsibly a construct this review tentatively terms Mathematical AI Literacy. The review concludes by outlining the theoretical contribution of this construct and the empirical gaps that remain, including the absence of standardized Mathematical AI Literacy instruments and the limited number of context-specific studies, particularly in Indonesia.]]>
