AI Accelerates Hacking Speed, Blockchain Protocols Need Continuous Audits, Says Cointelegraph Report

According to Cointelegraph on July 9, blockchain security experts warn that artificial intelligence is shortening the effective lifespan of smart contract audits, requiring continuous review instead of one-time assessments. TRM Labs policy head Ari Redbord stated that attack techniques are evolving faster than single audits can cover. CertiK reported $1.32 billion in stolen funds in the first half of 2026, with hackers increasingly using AI tools to exploit already-dormant DeFi protocol code repositories.

Recent cases highlight the risk: Aztec Connect was drained of $2.1 million three years after shutting down, while mySwap lost $300,000 six months post-closure. Security engineers at Zcash Shielded Labs discovered a four-year-old forgery vulnerability in the Orchard pool using a custom audit agent based on Claude Opus 4.8, underscoring that vulnerabilities remain exploitable long after launch.