News
Apple’s system update thanks AI for the first time: Claude and Codex teamed up to uncover multiple vulnerabilities. Apple is increasingly relying on AI companies to strengthen the security defense of the operating system.
2 min read
Source: Telegram AI频道
Apple’s system update thanks AI for the first time: Claude and Codex teamed up to uncover multiple vulnerabilities. Apple is increasingly relying on AI companies to strengthen the security defense of the operating system. In the latest system updates released for iOS, iPadOS, watchOS, tvOS, visionOS and macOS Tahoe, a large number of security vulnerabilities have been officially fixed, and a considerable part of the repair results are attributed to AI-assisted security research - this is the first time that Apple has directly mentioned the name of the AI model in the official CVE vulnerability attribution list. According to Apple's official security announcement, Anthropic's Claude model and OpenAI's Codex Security each helped discover three vulnerabilities, NVIDIA's AI red team contributed two, and Z.ai's GLM model also successfully helped fix one. Among them, security researchers Milad Nasr and Nicholas Carlini used the Claude model to help repair security vulnerabilities in key modules such as WebKit, WebKit Storage, and WebDAV. The Glasswing plan is initially showing results, with US$100 million to build an AI security alliance. Apple's in-depth application of AI security research is partly due to the "Glasswing Project" launched by Anthropic in April this year. The program costs $100 million and joins technology giants such as Apple, Microsoft, Google, Amazon AWS and NVIDIA to discover and patch vulnerabilities before potential attackers exploit them by obtaining the dedicated Anthropic "Claude Mythos Preview" model in advance. The core value of this model is to allow the AI model to play the role of "early warning" in security attack and defense, rather than waiting for vulnerabilities to be exploited by hackers before remediating them. Just as Microsoft announced two weeks ago that it will significantly increase the proportion of AI in Windows 11 security construction, Apple's list of vulnerability credits marks that the industry's top players are upgrading AI from an auxiliary tool to a core force in security governance. When AI can both discover and repair vulnerabilities, the balance of attack and defense in network security is quietly tilting. via AI News (author: AI Base)