AI × CryptoBullisharXiv – CS AI · 7h ago7/10
🤖
TRUST: A Framework for Decentralized AI Service v.0.1
Researchers introduce TRUST, a decentralized framework for auditing Large Reasoning Models and Multi-Agent Systems using hierarchical directed acyclic graphs, a causal attribution protocol, and multi-tier consensus mechanisms. The system achieves 72.4% accuracy in verification while maintaining privacy and preventing single points of failure, enabling tamper-proof auditing, leaderboards, and autonomous agent governance.