Security Firm Claims Exploit of Apple M5 Chip with AI Assistance
- A security firm developed a macOS kernel exploit targeting Apple’s M5 chip and Memory Integrity Enforcement (MIE) system.
- The exploit was created using Anthropic’s Claude Mythos AI, which helped identify bugs within a week.
- The exploit targets macOS version running on Apple M5 systems, starting from an unprivileged user account to escalate to root access.
- Anthropic’s Mythos AI is not publicly released and is available only to select companies and researchers under Project Glasswing.
The security firm Calif claims that their team used Anthropic’s Claude Mythos AI to bypass Apple’s new MIE protections in the M5 chip, highlighting the potential of pairing advanced AI models with human expertise in cybersecurity research.
This development underscores the ongoing challenges in maintaining robust security measures against rapidly advancing technology like AI-driven vulnerability identification and exploitation. (Source)
