AI crypto proxies have made all of DeFi insecure, says a security expert

Critics quickly hit back, arguing that the latest cryptocurrency hacks were caused by an operational failure instead.

Manuel Arauz, co-founder of smart contract security firm OpenZeppelin, went public on May 26 with an explicit recommendation that people should exit DeFi, all of it, including major corporations.

According to him, AI-powered crypto agents have successfully tilted the security game so far towards attackers that no protocol can currently be trusted to hold users’ funds.

Arauz warning

Software engineer books In a post on X;

“PSA: I now consider all DeFi services unsafe.”

He also said that he was privately advising friends and family to exit all DeFi sites, naming Aave, MakerDAO, and Compound as protocols he no longer considered safe.

His thinking is based on asymmetry: Defenders must find and fix every vulnerability, while attackers only need one vulnerability to cause damage. Now, with AI cryptographic agents able to scan smart contracts faster and more thoroughly than any human security team, Araoz feels that asymmetry has become impractical.

OpenZeppelin itself recently male And that cryptocurrency companies lost more than $3.4 billion due to hacks in 2025; However, it blamed most of this theft on compromised credentials, operational failures, and tokens shipped between audits, not on smart contract errors.

This year also saw a wave of attacks, amounting to more than $650 million Stolen In April alone. Of that amount, $292 million came from the KelpDAO exploit, with another $285 million pulled from Drift Protocol after what experts say were months of social engineering.

You may also like:

Rejection from X users

Against this background, Arauz’s warning was strong, but people responded immediately. One of those who criticized the post was Aave Chan Initiative founder Mark Zeller, who doesn’t hold anything back.

His counter was data-driven: he He pointed out Less than 10% of DeFi issues last year were caused by code-level vulnerabilities, with most failures, according to him, due to the vulnerability. Risk standards, poor collateral management, poor operational security, not AI-powered exploits.

Many others have echoed Zeller’s point, but with a slightly less warm tone. Phoenix Lab co-founder Sam McPherson Shown that premium DeFi platforms’ smart contracts were “pretty secure these days” and pointed to operational failures as the real reason behind most of the major hacks that have occurred recently.

Another X user, Polaris Finance developer Robert, made a similar distinction, Saying Actual smart contract exploits are “almost non-existent these days.” He added that recent breaches have largely involved centralized components that allow for human control rather than the static code underneath.

Ethereum co-founder Vitalik Buterin also has a different view on AI and its impact on cryptocurrency security, writing earlier this month that formal AI-assisted verification could make cryptocurrency systems more secure over time. According to him, developers can is used Artificial intelligence to write both the code and mathematical proofs of its correctness.