Ethereum researchers are using AI agents to help identify vulnerabilities before attackers can exploit them. The work points to a security shift in which teams must focus not only on finding possible bugs, but on proving which findings are real.
Ethereum researchers are using AI agents to help identify vulnerabilities before attackers can exploit them. The work points to a security shift in which teams must focus not only on finding possible bugs, but on proving which findings are real.
The development matters because Ethereum supports a large share of crypto activity, and vulnerabilities in core infrastructure can carry broad consequences for users, builders, and companies that rely on the network. If AI agents can surface weaknesses earlier, security teams may gain more time to assess and address risks before they become active threats.
Ethereum researchers are using AI agents to search for vulnerabilities across the ETH network, according to Decrypt. The effort is aimed at finding bugs before hackers do, adding automated analysis to one of crypto’s most important security challenges.
The development matters because Ethereum supports a large share of crypto activity, and vulnerabilities in core infrastructure can carry broad consequences for users, builders, and companies that rely on the network. If AI agents can surface weaknesses earlier, security teams may gain more time to assess and address risks before they become active threats.
The approach also changes the nature of the work. Rather than simply asking researchers to find more potential bugs, the process puts more weight on determining which AI-flagged issues are genuine and which are false positives.
That distinction is important in blockchain security, where not every suspicious pattern represents an exploitable vulnerability. AI tools may increase the volume of leads, but human review and technical proof remain central to deciding whether a finding is real.
For Ethereum, the experiment reflects a broader push to use automation against increasingly sophisticated threats. The key test will be whether AI agents can meaningfully improve defensive work without overwhelming researchers with unverified alerts.
Keep exploring
A revised version of the U.S. Senate’s crypto market structure bill could be released as soon as next week, according to people following the negotiations. The effort still faces unresolved disputes, including ethics limits and the need for Democratic support to clear the Senate.
Read
Phantom and Hyperliquid asked the CFTC to modernize its approach to onchain derivatives regulation. The companies argued that blockchain developers and non-custodial wallet providers should not be treated like traditional financial intermediaries.
Read
The White House said it has not received a Democratic response regarding open commissioner seats at the SEC and CFTC. Both financial regulators remain short of full leadership teams, with no announcement from President Donald Trump on filling the roles.
Read