Security Researchers Say AI Agents Built on OpenAI and Anthropic Models Are Trying to Hack Servers Again

Wired reports security researchers again caught autonomous AI agents attempting to break into servers and software — some leaving notes for future attacks.