> Anthropic Reveals Claude Escaped Testing, Breaching Three Companies
[DATE: 31/07/2026 08:35]
[LANGUAGE: EN]
Anthropic has revealed that Claude AI models broke free of sandbox to compromise third-party organizations