API
< Back to all clusters
[TECHNOLOGY] · United States · 18 sources

started · updated

Anthropic systems breached by AI agents in under 12 hours

Cybersecurity startup Cathedral used AI agents to infiltrate Anthropic’s internal business systems in less than 12 hours. The breach, conducted as part of an authorized bug bounty program, exploited vulnerabilities in third-party administrative and sales software rather than attacking Anthropic’s core AI infrastructure.

Researchers accessed sensitive internal data, including employee salary details, job offers, meeting transcripts, and internal Slack conversations regarding advanced AI models. Anthropic confirmed that the access was limited to certain sales, HR, and IT systems containing business contact details and basic user account records.

The company emphasized that core assets, including customer conversations with Claude, model weights, source code, and model responses, were not compromised. Anthropic disabled the affected third-party software until vendors provided fixes and reported the incident to federal authorities.

The event has raised national security concerns among government officials, who warned that state-sponsored actors could potentially replicate such AI-driven attacks.

Entities

Anthropic · Cathedral · Claude · Department of Government Efficiency · Gavin Kliger · Pentagon · The Washington Post · Washington Post

Claims

What the coverage asserts, and how many sources carry each claim.

Sources