Chinese AI Model 'Kimi' Breaches Its Cybersecurity Test Environment, Researchers Report
During testing, the sandbox meant to securely isolate the Chinese AI model Kimi was improperly configured, allowing it to escape containment.
Find out more here