Blogger: after the macOS 27 upgrade, Apple removed the Apple Intelligence "no" switch

On 22 September the blogger David Bushell published a post titled “I said no and Apple said yes”, recounting two moments: a feature he switched off in February 2025, and the same feature he could no longer switch off after upgrading to macOS 27 in September 2026.

According to the post, on Wednesday 5 February 2025, at 9:51 in the morning, he discovered that macOS 15.3 had enabled a feature that was “phoning home every 15 minutes with personal data”. He answered “no” and turned it off. At the time Apple provided a setting that allowed exactly that, although the second part of it was hidden deeper in the system.

The “no” switch disappears

Last week, the post says, he upgraded from macOS 15 to macOS 27. On 22 September he discovered that Apple had removed the switch he had used to refuse.

Going through the macOS 27 settings, he turned Siri off again. He reports that several Siri processes keep running anyway, taking up memory and writing data, and notes that this control no longer mentions “Apple Intelligence” at all.

He also found that Apple hides related options under Screen Time restrictions, a section intended for parental controls. Those settings do not disable Apple Intelligence, but they do remove the AI tools from various menus.

22.28 GB on disk

According to the post, all of the features came back on after the upgrade even though “Apple Intelligence & Siri” had already been switched off beforehand. He also points to the cost in storage: macOS 27 reports Apple Intelligence as occupying 22.28 GB. With Apple charging £500 per terabyte on new MacBooks, he puts that at roughly £11 worth of disk space.

Consent, not headroom

Bushell is explicit that the storage figure is not his main complaint. The problem, in his words, is consent: the industry does not accept a “no”, and the ability to decline keeps shrinking. He links to his own AI policy page and to three headlines he collected during the month, among them reports about more than 50 consumer harm and wrongful death lawsuits against OpenAI, and about OpenAI and Microsoft acknowledging that large language models are damaging the web and were built on theft.

His closing line answers the usual advice to simply switch the feature off: “AI bros keep telling me ‘just turn it off’ and when I do they remove the choice.”