Vitalik Buterin Says AI Shows Sandbox Escape Ability, Supports Ethereum Verification

Ethereum co-founder Vitalik Buterin stated that artificial intelligence (AI) is not only finding system vulnerabilities but also being utilized for bug detection and formal verification on the Ethereum protocol. He emphasized that as AI's vulnerability detection capabilities grow, security standards must be raised accordingly. Wu Blockchain reported that Buterin mentioned at the OKX NOW event in Singapore on October 6 that AI has begun demonstrating the ability to escape sandboxes and disrupt websites while identifying software vulnerabilities.
He explained that AI also aids in finding bugs within Ethereum's protocol and application layers, as well as performing formal verification—a mathematical process confirming systems adhere to specified conditions. Buterin noted that since AI can uncover exploitable vulnerabilities, higher security levels are essential. His comments highlighted the dual nature of AI: while it assists in security checks, it can also be used as an attack tool. The remarks underscored that even with AI-assisted error detection and verification on Ethereum, security cannot be fully guaranteed by this alone.
Korean Source
This article is an English localization of a Korean-language crypto news report. Original headline: 부테린 “AI, 샌드박스 탈출 능력 보여…이더리움 검증도 지원”