This study examined the growing intersection between artificial intelligence (AI) and systemic risk with attention to the open-source models as an emerging source of vulnerability in complex systems. The objective of the study is to systematically review and map global research trends on AI-driven systemic risk using a quantitative bibliometric approach. Data were sourced from the Scopus database, yielding a total of 218 relevant publications. The study employed VOS viewer to conduct keyword co-occurrence, overlay, network, and density visualisation, alongside descriptive analysis of publication trends, subject areas, funding sources, and country contributions.  The findings revealed a significant upward trend in publications, especially from 2020 onwards, indicating increasing scholarly interest in AI-related systemic risk. Network analysis identifies four major research clusters: i) machine learning and financial systems ii) decision-making, ethical technology and cybersecurity iii) risk assessment and deep learning iv) financial markets and electronic trading. Artificial intelligence emerges as the domain research hotspot, while areas such as ethical technology remain underexplored. The United States, China, and the United Kingdom are leading contributors, with computer science, social sciences, and finance as the most prominent subject areas. The study further highlights that while AI enhances innovation and efficiency, it simultaneously introduces systemic vulnerabilities through model opacity and security risks, especially within open-source ecosystems. The paper contributes to the literature by providing a comprehensive mapping of the intellectual structure of AI and systemic risk research. It also offers policy-relevant insights, emphasizing the need for robust governance frameworks.