Calif

A security startup named Calif claims researchers used Anthropic's AI model Claude Mythos to develop a macOS kernel exploit bypassing Apple's M5 chip protections, including Memory Integrity Enforcement (MIE), in under a week. The exploit grants root access from a local user account, highlighting concerns about AI accelerating the discovery of software vulnerabilities.

Researchers from security startup Calif exploited an Apple Mac M5 system using a preview version of Anthropic's Claude Mythos AI to build a macOS kernel exploit. This achievement demonstrates the potential capabilities of AI in cybersecurity. The exploit was made possible by the AI's ability to assist in building the kernel exploit.