Keeping your data from Apple is harder than expected
Scope and methodology of the study
- Study examined privacy controls for Apple default apps (Safari, Siri, iMessage, FaceTime, Find My, etc.) and found it is technically possible to limit data collection, but practically very hard for users.
- An experimental task where participants tried to prevent data sharing with Apple reportedly had a 0% success rate, even among relatively educated/technical users.
- Several commenters argue the researchers’ “surprise” is odd given years of dark patterns, confusing settings, and data monetization; others say experts are allowed to be surprised by just how bad UX can be in practice.
Siri, “learning from apps,” and unclear boundaries
- Major focus: Siri/Siri Suggestions “learn from app” behavior and how to disable it.
- By default, Siri can learn from app usage (including E2EE messengers) and share personal information across devices; opting out is per-app, not global, and confusing.
- Some say “you can just disable Siri”; others point out that:
- Voice control off ≠ learning off.
- Certain features (e.g., CarPlay) effectively require Siri to be enabled.
- Debate over what data leaves the device:
- One side: most Siri/Suggestions processing is on-device; data synced via end‑to‑end encryption; analytics are anonymized/differentially private and optionally sent.
- Other side: users can’t really verify this, settings are fragmented, and Apple could change behavior via updates.
Apple vs. other platforms on privacy
- Many still consider Apple “best of a bad bunch,” especially versus Google/Meta’s ad‑driven models.
- Others argue all major vendors are structurally similar: proprietary stacks, pervasive telemetry, and eventual exposure via sale, court orders, or compromise.
- Some cite research suggesting Apple is not clearly better than Google in all respects; others insist Android’s situation is likely worse but under‑analyzed.
Workarounds, alternatives, and threat models
- Suggested mitigations: disabling Siri/Suggestions via profiles (Apple Configurator/MDM), using tools like Little Snitch, or moving to privacy‑focused Android ROMs (GrapheneOS, CalyxOS) or even no smartphone at all.
- Counterpoints: custom ROMs have their own security/jank issues; hardware blobs remain; and many modern services push users toward mobile‑app dependence.
Critique of the article and presentation
- Several criticize the article’s diagram as confusing and its treatment of Touch ID/Face ID as if they were privacy problems despite being securely stored locally.
- Some view the university site’s heavy tracking/cookies as ironic but others call that a distraction from the Apple-specific issues.