The rapid evolution of Artificial Intelligence (AI) and the Social Industrial Internet of Things (SIIoT) has significantly impacted the agri-food supply chain, offering transformative solutions for security, efficiency, and sustainability. However, challenges related to data integrity, cyber threats, and system interoperability remain. This study provides a comprehensive analysis of AI-enabled secure SIIoT applications in the agri-food supply chain, addressing key security concerns and efficiency bottlenecks. It aims to develop a structured taxonomy of AI-driven security mechanisms, highlighting their roles in safeguarding SIIoT systems. A systematic literature review was conducted using reputable databases, including Google Scholar, ACM, DBLP, IEEE Xplore, SCOPUS, and Web of Science, focusing on peer-reviewed articles from the last six years. Additionally, multiple case studies were examined to validate the real-world application of AI-driven security frameworks in the agri-food industry. The findings indicate that AI-driven security solutions significantly enhance trust management, anomaly detection, and data privacy in SIIoT networks. The proposed taxonomy categorizes AI-enabled security mechanisms into five distinct areas, offering a structured reference for future research and practical implementations. Furthermore, case study analysis demonstrates the successful deployment of AI-driven security in real-world agri-food applications, emphasizing improved traceability and resilience against cyber threats. This study advances the field by identifying gaps in current research, proposing strategic recommendations, and outlining future directions for AI-enabled secure SIIoT systems in the agri-food research domain. The insights presented here provide a strong foundation for researchers, policymakers, and stakeholders in the agri-food sector to build more resilient and intelligent ecosystems.