Google's New AI Finds 55 Security Flaws in Its Web Engine

Google's New AI Finds 55 Security Flaws in Its Web Engine

Google built a security-focused AI called Gemini 3.5 Flash Cyber that found 55 confirmed security flaws in V8, the engine that runs JavaScript code in browsers. Ten flaws were found by no other AI tested. Rival Anthropic's model found 36. Only governments and trusted partners can use the tool, so outside experts have not confirmed Google's results.

Published

Read at another depth