CertiK reports AI agents autonomously tracing stolen crypto

CertiK: CertiK is a New York-based Web3 security firm founded by professors from Yale and Columbia universities that specializes in smart contract audits, formal verification, on-chain forensics, and AI-driven threat detection for blockchain protocols and decentralized applications. The company provides security tools and monitoring services to help protect crypto projects, exchanges, and DeFi platforms from exploits and compliance risks. In the news, CertiK released a report highlighting how agentic AI systems are now autonomously tracing stolen funds, drafting suspicious activity reports, and pausing smart contracts, while stressing that named human supervisors remain ultimately responsible for outcomes and governance. Web3 Applications: In blockchain contexts, these AI agents are applied to smart contract auditing, real-time transaction monitoring, and cross-chain fund tracing amid increasingly fast and complex attacks. AI Security Evolution: Agentic AI is advancing beyond simple flagging tools to act as autonomous investigators that reason through steps, use tools, gather evidence, and execute responses in live blockchain security and compliance environments. Accountability Requirements: Organizations must implement defined roles, bounded authority, audit trails, and human oversight when deploying AI agents for crypto security tasks to address governance and liability challenges.