I knew the pace of exploitation of security issues was getting faster and faster, but this chart shocked me. In just 5 years since 2021, the time between when a security issue is found and security researchers find people trying to exploit it went from 1.3 years to 1.6 DAYS.

While not entirely due to AI, I think hackers and exploiters are likely using AI to more quickly generate exploits. Project Glasswing was started because the upcoming version of Anthropic’s AI was finding zero-day exploits in browsers and operating systems at an alarming rate. Instead of just releasing the AI, they’re working with OS vendors from Apple to Microsoft to Linux to fix many of them before hackers take advantage of the holes.