AI-Augmented 6G Architecture for Trust-Centric Secure Consumer Transaction Processing in the Metaverse

The increasing reliance on digital transactions within the Metaverse introduces significant challenges related to security, fraud detection, and real-time risk management. Existing systems often lack the intelligence and responsiveness required to handle financial risks effectively across decentralized and immersive platforms. This paper presents an AI-augmented architecture that leverages sixth-generation (6G) wireless technologies to enable trust-centric transaction processing. The proposed framework integrates intelligent risk classification with ultra-low-latency communication and edge computing to support secure financial operations in virtual ecosystems. A curated dataset of 78,600 blockchain-based transaction records from the Open Metaverse was used to train and evaluate five machine learning models. The Gradient Boosting Classifier achieved the highest accuracy of 99.73%, followed by Random Forest at 98.36%, outperforming traditional models such as Support Vector Machine and Logistic Regression. The system also incorporates a continuous learning mechanism to adapt to evolving transaction behaviors. While the dataset provides comprehensive coverage of behavioral and transactional attributes, it is based on a simulated environment, which may limit generalizability. This work advances prior research by combining ensemble learning with distributed 6G processing, offering a scalable and real-time fraud detection solution for applications such as augmented reality commerce, virtual retail, and decentralized marketplaces.