AI Matters
Why it matters
Security firm's AI agents broke into Anthropic's internal systems
The Decoder AI agents hacked Anthropic, maker of arguably the most capable AI models for cyberdefense
Security startup Cathedral used AI agents to access Anthropic's sales, HR and IT data via third-party software in a bug bounty; Anthropic disputes how deep it went.
Why it matters
Even leading AI companies can be breached through outside software, which raises questions about how safe data held by AI providers is.
Latest brief · Oct 28, 2026