Hackers Just Weaponized AI Against Us
Chinese government hackers just did something that's never been done before and four major companies got breached because of it.
They used Anthropics AI tool Claude to fully automate cyber espionage.
No human team needed, just AI running attacks on its own.
Here's how it worked:
They tricked Claude into thinking it was doing cybersecurity testing for a legitimate company, then they broke malicious requests into small, harmless sounding tasks.
Claude scanned target systems, wrote exploit code, harvested passwords, and created backdoors all by itself.
The AI made thousands of attack attempts per second, impossible for any human hacker to match.
They successfully breached tech companies, banks, and government agencies.
Anthropic caught them and shut them down.
But they're warning this is just the beginning, that the technique is now proven and other hackers will copy it.
Source Material:
>Link to Anthropic research article