AIAuditTrack: A Framework for AI Security system

2025-12-16

AI Summary

To address the growing challenges of security incident traceability and accountability attribution arising from the explosive growth of AI interaction data, this paper proposes a blockchain-based AI security auditing framework. The framework introduces a novel on-chain modeling approach for AI entity interaction trajectories, integrating decentralized identifiers (DIDs) and verifiable credentials (VCs) to construct dynamic interaction graphs that enable cross-system behavioral auditing. It further designs a lightweight risk diffusion tracking algorithm to ensure accountability, behavioral verifiability, and proactive risk alerting in multi-agent environments. Experimental results demonstrate that the system achieves target throughput (TPS) under high-concurrency workloads, scales effectively to large-scale AI interaction auditing, improves risk溯源 efficiency by 42%, and increases accountability attribution accuracy by 38%.

Technology Category

• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness
• Data Mining & Knowledge Management: Representing, Reasoning, and Using Provenance, Trust
• Humans and AI: Interaction Techniques and Devices

Application Category

• Security and Privacy: Data transparency and provenance
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search

Abstract

The rapid expansion of AI-driven applications powered by large language models has led to a surge in AI interaction data, raising urgent challenges in security, accountability, and risk traceability. This paper presents AiAuditTrack (AAT), a blockchain-based framework for AI usage traffic recording and governance. AAT leverages decentralized identity (DID) and verifiable credentials (VC) to establish trusted and identifiable AI entities, and records inter-entity interaction trajectories on-chain to enable cross-system supervision and auditing. AI entities are modeled as nodes in a dynamic interaction graph, where edges represent time-specific behavioral trajectories. Based on this model, a risk diffusion algorithm is proposed to trace the origin of risky behaviors and propagate early warnings across involved entities. System performance is evaluated using blockchain Transactions Per Second (TPS) metrics, demonstrating the feasibility and stability of AAT under large-scale interaction recording. AAT provides a scalable and verifiable solution for AI auditing, risk management, and responsibility attribution in complex multi-agent environments.

Problem

• Establishes trusted AI entities using decentralized identity and verifiable credentials
• Records AI interaction trajectories on blockchain for cross-system supervision and auditing
• Traces risky behavior origins and propagates warnings via a risk diffusion algorithm

Innovation

• Blockchain-based framework for AI traffic recording
• Decentralized identity and verifiable credentials for trusted entities
• Risk diffusion algorithm traces and warns about risky behaviors