Apple has announced modifications to macOS privacy configurations designed to prevent third-party applications from scraping message records. The adjustment aims to curb potential misuse by software developers who leverage broad system permissions to ingest personal communication data into automated tools.
The policy revision follows public concern sparked by tech columnist Jason Aten, who reported receiving an unexpected notification from Meta’s newly launched general-purpose AI agent, Muse. The alert referenced a private conversation thread between Aten and a colleague held on Apple Messages. Aten stated he had not deliberately permitted the assistant to process his personal chats and expected those records to remain private.
The incident triggered widespread debate online regarding the security risks of granting artificial intelligence tools expansive access to local files, email accounts, and chat logs. Commentators likened these integrated agents to heavy machinery, noting that while automated assistants offer significant convenience, they pose distinct privacy hazards when granted excessive access to sensitive data stores.
Meta executive David Singleton pushed back against allegations of improper data collection, explaining the technical steps required for the integration to function. According to Singleton, Muse cannot read Apple Messages unless a user explicitly toggles on the Messages integration within the assistant's own menu and grants the application full-disk access at the macOS operating system level.
Despite Meta's explanation, Apple moved to adjust how operating system permissions protect communication databases. By tightening these boundaries, Apple intends to ensure that broad system privileges cannot be leveraged by third-party AI assistants to continuously monitor personal message streams.
