Apple Strips Out the Off Switch for On-Device AI in Latest macOS

A blogger recounts how Apple has steadily eroded user control over its AI features. In early 2025, macOS 15.3 quietly turned on a feature that transmitted personal data every 15 minutes; at the time, a toggle existed to shut it off, albeit buried in the settings. After upgrading (through Apple’s aggressively renamed version numbering) to macOS 27, the author found that the explicit opt-out had simply vanished, and that the AI tooling had re-enabled itself despite ‘Apple Intelligence & Siri’ already being switched off.

The practical annoyances pile up: Siri can be disabled in name only, since multiple unkillable background processes keep running and writing data for unclear reasons, and the install dumps roughly 22 GB of AI model data onto the disk — a real cost on machines where Apple charges a steep premium for storage. The only remaining way to hide some of the features is a workaround buried under Screen Time parental controls, which cleans up menus without actually disabling the underlying functionality.

The broader argument is about consent. The author frames the removal of the opt-out as emblematic of an AI industry that treats ‘no’ as a temporary obstacle rather than a decision to respect — advertising deepfake tools, re-enabling features across upgrades, and telling critics to ‘just turn it off’ while quietly deleting the switch that would let them. It’s an opinion piece rather than reporting, but it captures a growing user frustration with defaults that override explicit privacy choices.