Designing Trustworthy AI Architectures for Workforce Identity Verification in Distributed Recruitment Systems

The rapid digitization of recruitment systems has expanded the adversarial surface of workforce identity verification in distributed hiring environments. Virtual interviews, automated screening workflows, and AI-assisted assessments introduce new risks including deepfake substitution, proxy participation, credential spoofing, and synthetic identity fabrication. Traditional uthentication mechanisms are insufficient against AI-enabled deception across distributed system layers. This paper proposes a scalable and modular architecture for trustworthy AI-enabled workforce identity verification in distributed recruitment systems. The proposed model integrates identity authentication controls, behavioral anomaly detection, composite risk scoring, governance logging infrastructure, and human oversight interfaces within a hybrid microservices and event-driven architecture. A multi-dimensional threat taxonomy is introduced to formally characterize adversarial attack vectors, capability tiers, and system target layers. The architecture is evaluated under simulated adversarial conditions to assess detection accuracy, threat coverage, and latency performance. Results demonstrate high simulated detection effectiveness and low-latency performance under modeled attack distributions. By embedding fairness safeguards, explainability mechanisms, and governance traceability into system design, this work contributes a scalable reference architecture for secure and trustworthy identity verification in AI-enabled recruitment infrastructures. INDEX TERMS:Adversarial machine learning, distributed systems, identity verification, microservices architecture, trustworthy AI.