AI Agent Hackers Are Coming — A Turning Point in Cybersecurity

Anthropic warned that its new model Mythos can find and exploit vulnerabilities far faster than before.

Anthropic is allowing some organizations to pretest Mythos and has warned governments privately about attack risks.

Security leaders say AI agents can automate rapid, persistent exploitation, though AI still lags in contextual valuation.

Recent incidents show attackers used AI tools like Claude to breach hundreds of devices and steal government data.