New macOS vulnerabilities were exposed by Anthropic’s Mythos: report
9to5Mac • • technology
Key Points:
- Researchers from Calif, a Palo Alto-based security firm, used Anthropic's Mythos AI model to identify macOS security vulnerabilities that enable privilege escalation exploits by corrupting the Mac's memory.
- The exploit discovered links two bugs and several techniques to gain unauthorized access to protected parts of the device, potentially allowing hackers to seize control if combined with other attacks.
- Calif delivered a detailed 55-page report to Apple, which is currently reviewing the findings and prioritizing security to address the potential vulnerabilities.
- Calif's CEO emphasized that the exploit required both Mythos AI and human cybersecurity expertise, and the company plans to publicly release details after Apple resolves the issues.
- The vulnerabilities are expected to be fixed promptly, according to Calif's team, underscoring ongoing efforts to strengthen macOS security.