The Rise of AI Agents in Securing Smart Contracts
In 2025, the world witnessed a significant shift in how digital assets are protected. Developers are now leveraging artificial intelligence (AI) agents to safeguard smart contracts that control billions of dollars in digital assets. This move comes in response to a surge in crypto thefts, with hackers managing to steal over $3.4 billion from blockchain platforms during the year.
The losses in 2025 were not spread out across numerous small incidents but were instead concentrated in a few major breaches. These three large-scale events accounted for nearly 70% of the total value stolen. One of the most notable cases was the Bybit exchange hack, which resulted in the loss of approximately $1.4 billion — one of the largest cryptocurrency thefts ever recorded.
AI Agents as Security Tools
OpenAI is collaborating with Paradigm and OtterSec to explore the potential of AI agents in detecting vulnerabilities within real blockchain environments using its EVMbench framework. These AI agents are designed to review smart contracts and identify and fix security issues that can lead to significant financial losses.
Smart contracts are automated programs that manage more than $100 billion in open-source digital assets. Any error in their code can have real-world consequences, affecting both large and small investors. The increasing frequency and scale of attacks have made it clear that relying solely on human audits is no longer sufficient. Live contracts face new and evolving threats that may not be present during the audit process, making manual reviews time-consuming and costly.
Continuous Monitoring with AI
Instead of waiting for the next manual audit cycle, which may come too late to prevent an attack, developers are turning to AI agents for continuous monitoring of live smart contracts. These agents can detect hidden code irregularities much faster than humans, who may need days or even weeks to identify issues.
EVMbench uses AI agents in test environments to help developers understand how smart contracts may perform under real-world conditions before deployment. The agents first detect vulnerabilities, fix them without breaking the contract’s functionality, and then attempt to exploit the weakness to drain funds if the problem persists.
The Double-Edged Sword of AI
According to early results, AI agents are better at exploiting vulnerabilities than they are at safely fixing them. This has raised concerns among experts that hackers could misuse AI-powered tools to exploit weaknesses in blockchain systems more efficiently than ever before.
Machines are learning to break into weak contracts faster than ever before, as current AI agent systems now succeed in exploiting more than 70% of vulnerabilities compared to earlier models with less than a 20% success rate. Attackers are moving away from manual hacking methods and toward AI agents that can scan large amounts of code and test different attack paths without direct human input.
Future Implications
As this trend continues, experts predict that AI agents will soon be able to move funds, approve transactions, and manage financial tasks automatically on behalf of users. American technologist Jeremy Allaire has stated that billions of AI agents will soon use stablecoins to send and receive payments across blockchain networks. Changpeng Zhao (CZ), founder and former CEO of Binance, also believes that crypto could become the native payment layer for AI-driven systems in the future.
These developments make AI agents increasingly useful for both users and attackers, as they will soon interact directly with contracts in real financial environments where actual money is at stake.
Industry Concerns and User Safety
Industry leaders have raised concerns about user safety. Managing partner at Dragonfly, Haseeb Qureshi, warned that many users still worry about sending funds to the wrong address or approving a harmful transaction by mistake through crypto transactions. To address this issue, Qureshi proposed that AI-operated wallets could soon interact with the blockchain without users needing to understand the complex processes involved.
In this way, AI agents can assist in reducing human errors in audits and in protecting smart contracts by continuously monitoring systems. However, they can also increase the rate at which attackers discover vulnerabilities in the system, enabling exploits to scale much faster.
This creates a security issue where AI systems developed to protect decentralized finance platforms could also be the most effective at attacking them if they fall into the wrong hands.