Apple to Enhance Mac Privacy Controls Against AI Agents Following Meta Muse Concerns

Apple has announced plans to implement new privacy controls on Mac computers to safeguard users against artificial intelligence agents, following user complaints that Meta Platforms' Muse may have accessed private messages. The upcoming guardrails are designed to ensure that users granting an app full disk access can only do so through explicit action, according to the company's developer site.

Full disk access allows apps to reach sensitive information including files, emails, messages, and browsing history, while also enabling backup apps to function properly on Macs. Apple noted that this feature largely bypasses developer application programming interfaces intended to protect private data. The new safety measures will apply to third-party apps as well, though California-based Apple did not specify when the rollout will begin.

Apple stated that some developers are utilizing full disk access in ways that expose entire computer systems without the user's full understanding, which can also compromise the privacy of individuals communicating with the user. The company acknowledged that autonomous AI agents are growing increasingly capable, making the associated risks substantially higher.

This initiative follows recent challenges for Meta's AI projects. Meta introduced Muse last month as a personal AI agent, but user complaints soon emerged claiming the tool could access private information such as messages. In July, Meta also faced backlash and scrapped Muse Image after it was automatically linked to Instagram accounts without user consent.