LOADING...

Claude AI aces cybersecurity challenges, sparks debate

Technology

Anthropic's AI, Claude, is turning heads at student cybersecurity competitions.
At Carnegie Mellon's PicoCTF, it cracked 11 out of 20 hacking tasks in just 10 minutes and landed in the top 3%—all with barely any human help.

AI's dual role in cybersecurity

Claude isn't just a one-hit wonder. It also nailed 19 out of 20 challenges at Hack the Box, something only a handful of human teams have pulled off.
Even DARPA-backed AIs like Xbow are topping bug bounty charts solo.
Experts say it's time to rethink cybersecurity—AI might soon be both our best defender and toughest rival online.