AI Threatens DeFi: Expert Warns of Unsafe Crypto Space

Manuel Aráoz, a co-founder of OpenZeppelin, has publicly stated that the decentralized finance (DeFi) sector is unsafe, citing the escalating capabilities of artificial intelligence in identifying vulnerabilities. He has personally advised friends and family to withdraw from all DeFi positions, including those considered stable. Aráoz emphasizes the asymmetric nature of smart contract security, where attackers need only one exploit to succeed, while defenders must fix every flaw.

Recent developments, particularly with AI models like Anthropic's Mythos, have demonstrated an unprecedented ability to uncover critical bugs in established software. This has heightened concerns within the crypto industry, with major exchanges reportedly seeking access to such powerful AI tools. The trend of escalating hacks, with April marking a record month for incidents, underscores these security challenges.

While some incidents, like the StablR compromise, involve centralized attack vectors such as compromised multisignature wallets, the overarching concern about AI's role in uncovering code flaws persists. Experts like Uttam Singh advocate for circuit breakers and emergency halt powers. However, others, such as Marc Zeller, argue that most DeFi issues stem from non-code-related factors, and that defensive applications of AI will ultimately enhance protocol resilience.