Apple has announced plans to introduce additional safeguards for artificial intelligence (AI) tools seeking access to data stored on Mac computers, following complaints over Meta’s AI agent Muse.
Muse is designed to perform complex tasks on behalf of users, including cancelling unwanted subscriptions and negotiating better prices. However, some users have raised concerns that the AI agent may access private messages without their knowledge.
Unlike iPhones and iPads, where apps are generally isolated from one another through a security mechanism known as sandboxing, Mac computers offer a Full Disk Access feature. With user permission, the feature allows applications such as cloud backup services to access data across the device.
Apple said some developers were using Full Disk Access in ways that could potentially expose users to privacy and security risks.
The company said it would introduce additional controls to ensure that users who choose to grant apps such extensive access must take explicit action before doing so.
Apple also warned that the risks associated with broad data access could increase as AI agents become more capable and autonomous. The company said users should clearly understand the potential risks before allowing AI systems to access their data.
The announcement followed allegations by technology columnist Jason Aten, who claimed that Meta’s Muse had accessed private messages on his Mac without his permission. Aten said he had not enabled Full Disk Access.
Meta spokesperson Andy Stone rejected the claim, saying Muse can access content from Apple’s Messages app only when users explicitly enable both Full Disk Access and the Messages connector.
Stone said the access can also be revoked by users at any time.
Apple declined to comment beyond its published statement, while Meta did not immediately respond to requests for further comment.
