It would be a very useful feature for Apple to do the same for contacts: the app would think it's getting access to your contacts, but would only actually receive a subset of them, and be none the wiser. This would be a tremendous boon for privacy.
What we really need to see from Apple is a permissions index in the app store that allows me to inspect, and consider, the permissions that an app will request before installing that app.
I shouldn't have to install the app (or do laborious research online) to discover what permissions it will attempt to utilize and which of them are required to function.
It would be trivially easy to list that in the app store, for each app.
One of the examples given here was an app that gives you safety alerts. A navigation app might give you useless directions. There are a thousand ways this could go horribly wrong.
I suppose iOS could present some warning, but that might interfere with the UI, or be misunderstood.
Or even a selected subset of real+fake data. (With the ability to define sets that could be reused across application.)
Sometimes I want the app to have access to some of my contacts, but not all of them. For example, my work contacts but not my personal friends, or vice versa. Or simply "my contacts that told me they were also using this app". PLUS hundred fake contacts to poison the data.
It's not - that's a violation of the App Store TOS. That's also not what's happening here - you can use clubhouse without allowing contacts access, but you can't invite someone to the closed beta without allowing it.
The policy solution clearly doesn't work in all scenarios because Clubhouse is still on the store. But an on-they-fly permission model that allowed the user to deny the permission invisibly or share a subset of their contacts would completely solve the problem regardless of whether or not Apple was effective at moderating.
Apple could still do whatever moderation they wanted to reduce annoyances for the end user, but the sandboxing approach would catch any apps they missed or refused to moderate.
This would also solve the problem where an app legitimately needs some access to contacts to run, but doesn't need access to the entire list. Clubhouse does need access to some contacts to invite someone to the beta, but it does not need access to the entire contacts list, and there's no reason for it to have the ability to tell whether or not a user is providing the full list.
not if the app still functions with deliberately reduced functionality. What i want to have is for the app to be unable to tell the difference between being denied permission, and having no data (or be sent fake data).
Actually I would take it further and say that I should be able to define its response or have it render a random but plausible template response. Otherwise a zero len response is too obvious that you didn't give it permissions.
I once had an app yell at me for not giving my GPS permission, but then yell at me again when I enabled a mock GPS on Android. It really shouldn't have been able to know I was mocking location.
On the iOS App Store, none of the optional permissions can be required for an app to perform it's basic functions - that's a store policy, and it's generally well enforced. Obviously if your app's function is mapping, GPS can be required to use those features (but only at the user's discretion - ie while running or all the time, granular or coarse), but the app can't just refuse to launch without it.
https://en.wikipedia.org/wiki/Citizen_(app)
> Citizen is a mobile app that sends users location-based safety alerts in real time.[1][2][3][4] It allows users to read updates about ongoing reports, broadcast live video, and leave comments.[1][2] The app uses radio antennas installed in major cities to monitor 911 communications,[5] with employees filtering the audio to generate alerts.[5] In March 2020, Citizen added the COVID-19 digital contact tracer SafePass. The app is currently available for iOS and Android devices[6] in 20 cities,[7] including New York City, the San Francisco Bay Area, Baltimore, Los Angeles,[8] Philadelphia.[9] Detroit,[10] Indianapolis,[11] Phoenix,[12][13] Cincinnati,[14] Chicago, Minneapolis, Saint Paul, and Cleveland.[15]
Most iPhone chat apps for example work perfectly fine with zero permissions granted yet provide the option to send pictures, invite contacts, use mic/camera, send gps location, etc if a user is so inclined. With a permissions index, you would likely end up with the majority of apps listing all permissions and users would simply ignore it.