As outlined in TechCrunch, Apple has introduced significant security updates to its macOS platform to mitigate risks associated with the rise of artificial intelligence agents. These autonomous tools often require extensive system permissions, prompting the company to refine how users manage access to sensitive data.
The update, arriving with macOS 27.3, addresses concerns regarding AI agents that operate directly on user devices. Unlike standard chatbots, these agents can execute actions across files and applications. Apple is splitting existing security controls into two distinct settings: System & App Protection Override and Accessibility. This change aims to clarify the implications of granting broad permissions. Furthermore, Apple is implementing a transparency indicator and providing users with granular control, allowing them to require authorization each time an app requests access. These measures follow reports of potential vulnerabilities in third-party AI applications. While the update does not block the use of AI agents, it mandates that developers adapt their software to handle restricted access more effectively. Developers requiring full disk access for legitimate functions, such as backups, may apply for specific entitlements. By providing increased visibility and more nuanced configuration options, Apple seeks to balance the growing demand for local AI model execution with the necessity of maintaining robust user privacy and system integrity.
Source: TechCrunch
