Apple is introducing additional controls around the Full Disk Access setting in macOS, citing growing risks associated with AI agents that can access files and other sensitive information on a user’s computer.
The announcement comes just days after a journalist claimed that Meta’s Muse app on Mac had accessed his private messages, a claim Meta disputed. Apple said the Full Disk Access feature was originally designed to ensure that backup applications could work properly, but the growing capabilities of AI agents have increased “the risks associated with this level of access.”
Apple’s statement followed a report by Inc. columnist Jason Aten, who said Meta’s Muse AI agent appeared to know the contents of his private messages even though he claimed he had never given the agent permission to access them. The report raised broader questions about security and user trust when desktop-based AI agents are given the ability to control applications, access files and interact with private messages.
Apple’s decision to introduce additional restrictions also follows a Wired report that highlighted a flaw in ChatGPT’s Mac application that could potentially have allowed hackers to obtain sensitive information.
Desktop AI agents can require users to provide their applications with expanded access to files, messages and other personal information stored on their computers. Users can grant some of these permissions by changing settings within macOS.
In the case of Muse, users can optionally enable Full Disk Access. Apple explains that this permission can give an application access to files, mail, messages and even browsing history.
“Some developers are using Full Disk Access in ways that could put users at risk, exposing everything on their systems…without users’ full knowledge and understanding,” Apple said in a new blog post directed at developers.
Apple said it will introduce additional controls going forward to ensure that users who genuinely want to give an application this unusually broad level of access must take “very explicit user action” before granting the permission.
“Addressing this is critical. As AI agents become increasingly capable and autonomous, the risks associated with this level of access will grow substantially,” Apple said. “We are committed to ensuring users clearly understand these risks before granting such access, so they can make informed decisions about their own data and privacy.”
