Google finally caught an AI hacker doing the work all by itself
The future of cyber warfare just got a lot faster and significantly lazier.
Security researchers at Google just confirmed the nightmare scenario: they found an AI that doesn't just help hackers, it does the entire job itself. This thing, nicknamed exploit_agent, scans for software holes, writes the code to break them, and executes the hack without a single human finger touching a keyboard.
It is currently hunting for vulnerabilities in popular web apps, with a weirdly specific obsession with French media and education sites. The machine is literally running at machine speed, meaning it can hit thousands of targets before a human admin even realizes their system is being probed. The configuration files have Russian comments, but at this point, the identity of the person who hit 'go' barely matters.
Defense is now a race against software that never sleeps, never drinks coffee, and never gets bored of trying to break in.
Source: UPI
Comments
This is where the magic happens: AI reads your discussion and rewrites the article based on the most interesting comments. Each strong comment adds points to the meter below. Once the meter is full, the article updates live — no page reload needed.