Security Researcher Warns AI Coding Agents Have Made DeFi Unsafe
DeFiSecurity
Bearish

Security Researcher Warns AI Coding Agents Have Made DeFi Unsafe

Manuel Aráoz, a top cryptocurrency security executive, warned that AI-powered coding agents have introduced new vulnerabilities to DeFi protocols. Aráoz said he has advised contacts to exit even major DeFi positions over concerns about smart contract safety.

May 28, 2026, 12:12 PM1 min read

Key Takeaways

  • 1## Warning on AI-Assisted Code Manuel Aráoz, a prominent figure in cryptocurrency security, raised concerns that AI-powered coding agents have fundamentally altered the threat landscape for decentralized finance.
  • 2Aráoz did not specify which protocols or coding practices pose the highest risk, but argued the shift toward machine-generated contract code creates attack surface that traditional auditing practices may not catch.
  • 3## Personal Risk Assessment Aráoz said he has recommended to friends and family that they reduce or exit positions in major DeFi protocols.
  • 4The scope of his concern extends beyond obscure projects to established platforms, suggesting he views the risk as sector-wide rather than isolated to smaller or newer projects.
  • 5## Open Questions The warning lacks specifics about which AI tools, coding patterns, or DeFi categories pose the greatest danger.

Warning on AI-Assisted Code

Manuel Aráoz, a prominent figure in cryptocurrency security, raised concerns that AI-powered coding agents have fundamentally altered the threat landscape for decentralized finance. Aráoz did not specify which protocols or coding practices pose the highest risk, but argued the shift toward machine-generated contract code creates attack surface that traditional auditing practices may not catch.

Personal Risk Assessment

Aráoz said he has recommended to friends and family that they reduce or exit positions in major DeFi protocols. The scope of his concern extends beyond obscure projects to established platforms, suggesting he views the risk as sector-wide rather than isolated to smaller or newer projects.

Open Questions

The warning lacks specifics about which AI tools, coding patterns, or DeFi categories pose the greatest danger. No on-chain data or published audit findings have been cited to support the claim, and Aráoz has not released detailed technical documentation of the vulnerabilities he fears.

Why It Matters

For Traders

If AI-generated smart contract code carries undetected vulnerabilities, DeFi platforms face heightened exploit risk; unconfirmed warnings like this typically precede deeper technical scrutiny.

For Investors

Structural concerns about smart contract safety could reshape risk premium across DeFi tokens and trigger broader due-diligence reviews of protocol engineering practices.

For Builders

If AI-assisted development becomes a liability vector, teams may need to adopt stricter verification and testing protocols, raising engineering and audit costs.

Sources

Related Articles

Latest News