Abstract
In collaborative systems, the effective completion of tasks hinges ontask-specific trust evaluations of potential devices for distributedcollaboration. However, the complexity of tasks, the spatiotemporal dynamism ofdistributed device resources, and the inevitable assessment overheaddramatically increase the complexity and resource consumption of the trustevaluation process. As a result, ill-timed or overly frequent trust evaluationscan reduce utilization rate of constrained resources, negatively affectingcollaborative task execution. To address this challenge, this paper proposes anautonomous trust orchestration method based on a new concept of semanticchain-of-trust. Our technique employs agentic AI and hypergraph to establishand maintain trust relationships among devices. By leveraging its strengths inautonomous perception, task decomposition, and semantic reasoning, we proposeagentic AI to perceive device states and autonomously perform trust evaluationsof collaborators based on historical performance data only during device idleperiods, thereby enabling efficient utilization of distributed resources. Inaddition, agentic AI performs task-specific trust evaluations on collaboratorresources by analyzing the alignment between resource capabilities and taskrequirements. Moreover, by maintaining a trust hypergraph embedded with trustsemantics for each device, agentic AI enables hierarchical management ofcollaborators and identifies collaborators requiring trust evaluation based ontrust semantics, thereby achieving a balance between overhead and trustaccuracy. Furthermore, local trust hypergraphs from multiple devices can bechained together to support multi-hop collaboration, enabling efficientcoordination in large-scale systems. Experimental results demonstrate that theproposed method achieves resource-efficient trust evaluation.