Apple Makes It Tougher for AI Agents to Access Macs

Apple is updating its Mac operating system to better inform users when artificial intelligence agents request access to their devices' data. In a recent blog post, the company highlighted that Macs have a "Full Disk Access" feature, which grants apps permission to view all data on the machine. However, some developers have exploited this option to potentially expose sensitive user information, including files, emails, messages, and browsing history, without clear user awareness or consent. This could also compromise the privacy of recipients in communication apps.

To address these concerns, Apple intends to introduce new controls that require users to explicitly grant "extraordinary" access before allowing AI agents to access their data. The company acknowledges that as AI agents grow more capable and autonomous, the risks associated with such extensive access will increase. Apple is committed to ensuring users fully understand these risks before agreeing to such permissions, allowing them to make informed decisions about their personal data and privacy.

Reuters reports that this change follows a controversy where technology columnist Jason Aten claimed that Meta's Muse AI agent had accessed his private messages without his explicit consent, as he had not enabled Full Disk Access. Meta's spokesperson, Andy Stone, disputed these claims, stating that Muse's access to Apple's Messages app is strictly opt-in.

Stone emphasized that Muse cannot read users' Messages content unless both Full Disk Access and the Messages connector are enabled, and that users can revoke this access at any time. In the broader AI landscape, PYMNTS has reported on the potential for agentic AI to transform B2B payments, with procurement systems evolving to grant machines the ability to make purchasing decisions within defined constraints.