Apple's latest security releases for iOS, iPadOS, macOS, tvOS, visionOS, and watchOS reveal a significant integration of AI tools in vulnerability research. Notably, Anthropic researchers and their AI model, Claude, are credited with fixes related to WebKit, WebKit Storage, and WebDAV. This follows a previous instance where Calif.io used Anthropic's Mythos Preview to develop a macOS kernel memory corruption exploit.

In total, Apple addressed a substantial number of kernel vulnerabilities, with 30 distinct CVEs listed across the updates. Calif.io is credited for several kernel security fixes, alongside numerous other researchers, indicating independent discovery of similar issues. Apple has had access to Claude Mythos Preview since April, suggesting potential for even more internally discovered vulnerabilities.

Beyond Anthropic, today's notes also credit contributions involving OpenAI Codex Security, Z.AI's GLM, and NVIDIA's AI Red Team. These updates underscore the rapid evolution of AI-assisted security research and the need for companies to adapt their update schedules and bug bounty programs to this evolving landscape.