Autonomous commerce agents get improved fraud detection and control

Securing Agentic Commerce Through Behavior-Aware Fraud Detection and Cybersecurity for Trusted AI-Agent Payment Authorization

Artificial Intelligence

Summary

Digital systems using independent AI agents to make payments can be tricked by fake agents or compromised systems, leading to unauthorized spending. The researchers designed a system called ACAGS that uses machine learning to watch how agents behave and check permissions to spot fraud and misuse. They tested this system using made-up financial data and showed it could detect impersonation, unusual behaviors, and improper authorization with high accuracy. The system also explains why it flags something suspicious to help people understand and trust its decisions. However, more testing with real-world data is needed to confirm its effectiveness.

What this means in practice

Tested on simulated data.

Authors

Kayode L. Ogunsusi

Abstract

The growing use of autonomous artificial intelligence agents in digital commerce creates new security concerns involving agent impersonation, unauthorized payments, compromised agents, and misuse of delegated authority. This research introduces the Autonomous Commerce Agent Guard System (ACAGS), a cybersecurity framework combining supervised machine learning, behavioral analytics, agent-specific authorization controls, and explainable AI. The framework was evaluated using synthetic financial transaction data augmented with simulated agent characteristics and five adversarial scenarios. XGBoost was selected following comparisons with alternative classifiers. Across ten experimental runs, behavioral features improved mean F1-score from 52.12% to 67.49%, with agent-specific features increasing performance to 69.46%. The framework achieved detection rates of 99.77% for abnormal agent behavior, 99.73% for impersonation, and 100% for malicious delegation and authorization-scope violations. Compromised-agent detection reached 72.27%. ACAGS also supported APPROVE, REVIEW, and REJECT authorization outcomes, with SHAP explanations providing model interpretability. The results demonstrate the potential of combining behavioral intelligence and delegated authorization controls to strengthen simulated agentic-commerce payment security, while highlighting the need for validation using real-world transactions.
View PDFOpen arXiv