OpenAI is diving further into the crypto security space with the launch of EVMbench, a new testing framework designed to measure artificial intelligence’s understanding and potential security of smart contracts on Ethereum and similar blockchains.
Smart contracts, self-executing code deployed on blockchains such as Ethereum, support decentralized exchanges, lending protocols and a wide range of on-chain financial applications. Since these contracts are typically immutable once deployed, vulnerabilities can be severe.
EVMbench is OpenAI’s attempt to understand whether modern artificial intelligence systems are up to the task of helping prevent these problems. The benchmark, built in partnership with cryptocurrency investment firm Paradigm, exploits real-world smart contract vulnerabilities previously discovered through audits and security competitions.
The system measures performance on three core capabilities: identifying security bugs, exploiting those bugs in a controlled environment, and fixing vulnerable code without breaking the contract.
OpenAI said its goal is to establish a clear standard for evaluating artificial intelligence systems in blockchain security, especially as decentralized finance continues to secure billions of dollars in user funds. The risks of smart contracts are only getting bigger.
“Smart contracts typically secure over $100B worth of open source crypto assets. As AI agents advance in their ability to read, write, and execute code, it is increasingly important to measure their capabilities in an economically meaningful context and encourage the defensive use of AI systems to audit and strengthen deployed contracts,” OpenAI wrote in a blog post.
Read more: Most Influential: Sam Altman
