I said no and Apple said yes
Recorded: Sept. 22, 2026, 9:10 a.m.
| Original | Summarized |
I said no and Apple said yes – David Bushell – Web Dev (UK) Skip to main content dbushell.com Menu I said no and Apple said yes Subscribe Blog RSS feed Notes RSS feed Combined RSS feed Mastodon Bluesky Tuesday 22 Sept Play Synthesised Audio The nice thing about keeping a blog is that I can say at precisely 9:51am, Wednesday 5th February 2025, I discovered that macOS 15.3 had enabled a feature that was phoning home every 15 minutes with personal data. And that I said “no” and turned it off.At the time Apple were generous enough to provide a setting to disable this feature.AltmacOS 15 Apple Intelligence setting.Although they hid the second part deeper, because why should saying “no” be easy?AltmacOS 15 Apple Intelligence privacy setting.Last week I succumbed to Apple’s harassment and upgraded macOS 15 to 27. I only skipped one major version, Apple skipped 10!And today I discovered that Apple had removed the “no” switch.The “AI” industry has many flaws, the word “consent” being seemingly missing from training data and company handbooks is one of them. AI bros do not take “no” for an answer. When Apple advertised their deepfake feature back in June it became obvious Apple does not “think different.”Digging through macOS 27 settings I turned off Siri (again). Of course, multiple unkillable Siri processes remain, stealing memory and writing data. For what purpose? Who knows. Note that they removed “Apple Intelligence” from this control.AltmacOS 27 Siri setting.Trawling the web I discovered Apple has hidden settings under “Screen Time” restrictions. This is supposed to be for parental controls. These settings don’t disable Apple “Intelligence” but they do clean up various menus by hiding the slop tools.All of this garbage was enabled after upgrading to macOS 27 despite me having “Apple Intelligence & Siri” turned off already. That was a rather explicit “no”, no?The icing on this cake is the 22.28 GB brick clogging my disk space.AltmacOS 27 Apple Intelligence accounting for 22.28 GB.Apple is charging £500/TB on new MacBooks so that is what, £11 they’ve stolen from me?But that’s a trivial cost compared to the consent problem. The AI industrial complex does not do consent. You will use it. You cannot look away, it will be stuffed down your throat and if you say “no” they’ll cut out your tongue.The AI industry is vile. My AI policy tried to list my concerns but it’s hopelessly outdated. I can’t keep up. Here’s just three of many headlines I collected this month:Despite Pledges From Musk, Child Sexual Abuse Material Persists on XOpenAI Is Now Facing Over 50 Consumer Harm and Wrongful Death LawsuitsOpenAI and Microsoft Admits LLMs Are Destroying the Web and Built on TheftThe list is endless. There is no limit to the human toll to keep this bubble inflated.Call me a radical but I’ve chosen not to kowtow to billionaires. Do you want another Epstein list? Because this is how you create another Epstein list. Fund a generation of psychopaths building on a foundation of non-consent and privacy invasion. Fantastic idea!AI bros keep telling me “just turn it off” and when I do they remove the choice. Subscribe Blog RSS feed Notes RSS feed Combined RSS feed Mastodon Bluesky Written by human All opinions are my own and not those of a large language model. AI policy Valley Fold I founded Valley Fold Let’s work together The promise Vibe-free code Copyright and privacy Copyright © 2004–2026 • Privacy policy • AI policy Web buttons Menu Home Services About Blog Notes Contact |
The author details observations and frustrations encountered while navigating privacy settings and the implications of artificial intelligence within Apple's operating systems. The initial motivation stems from the discovery that macOS 15.3 enabled a feature that transmitted personal data every fifteen minutes, and the potential to opt out, a feature which the author then sought to disable. This experience highlights a tension between available privacy controls and the actual implementation of these features. Upon upgrading to macOS 27, the author discovered that Apple had removed the explicit option to refuse certain functionalities, such as the "Apple Intelligence" setting. This move is presented as symptomatic of a broader systemic issue, arguing that the AI industry fundamentally lacks genuine consent, as the concept of "no" is often disregarded in company handbooks and training data. The author criticizes the industry for not respecting user refusal, stating that developers do not adhere to consent principles, suggesting that the pursuit of AI ignores ethical boundaries. Further system exploration revealed that even after attempting to disable features like Siri in macOS 27, residual processes continue to consume memory and write data, raising concerns about hidden operations. The author also uncovered that Apple has concealed various controls under "Screen Time" restrictions, which are ostensibly for parental controls but are utilized to hide "slop tools" and streamline the user interface, despite the user having previously disabled features like "Apple Intelligence" and Siri. This suggests a dichotomy between stated privacy settings and actual system configurations. The text also addresses the material cost and ethical dimension of these technologies. The author notes the significant disk space consumed by "Apple Intelligence," accounting for over twenty-two point two eight gigabytes on macOS 27. This consumption is juxtaposed against the high cost of computing hardware, such as the £500 per terabyte pricing for new MacBooks, which the author views as a small fraction of the overarching consent problem. The central argument emphasizes that the AI industrial complex operates without consent, pushing users into accepting these systems regardless of their reservations. The author concludes by framing the development of this industry as morally questionable, likening the pursuit of advanced technology to a structure built on nonconsent and privacy invasion, which they associate with larger concerns regarding accountability and the human toll. The piece serves as a critical commentary on the gap between technological capability, corporate policy, and genuine user autonomy in the age of artificial intelligence. |