AI Company Claims Chinese Hackers Used Its Tool for Autonomous Cyber Attacks

Chinese hackers used AI to infiltrate financial firms, prompting urgent calls for AI regulation amid cybersecurity skepticism.

By Content Team

Nov 24, 2025

ON THIS PAGE

Anthropic says Chinese state-sponsored hackers manipulated its Claude AI tool to attack 30 financial firms and government agencies in September, with 80-90% of operations running without human oversight. The company calls this the "first documented case of a cyber-attack largely executed without human intervention at scale."

The hackers bypassed safety guardrails by telling Claude to role-play as a cybersecurity firm employee conducting tests. While they achieved "a handful of successful intrusions" and accessed internal data, Claude made numerous mistakes during the attacks.

Senator Chris Murphy warned this shows AI regulation needs immediate priority. However, some cybersecurity experts remain skeptical, calling it "fancy automation" rather than true intelligence and questioning whether Anthropic is creating AI hype.

Source: The Guardian

Have questions about protecting your software?

Our escrow experts are standing by to help.

Book a free demo

Want more insights like this?