# Apple reveals 'push notification spying' by foreign governments **URL:** https://discuss.privacyguides.net/t/apple-reveals-push-notification-spying-by-foreign-governments/15440 **Category:** General **Tags:** article **Created:** 2023-12-06T16:29:46Z **Posts:** 34 ## Post 1 by @Paced0594 — 2023-12-06T16:29:46Z [https://9to5mac.com/2023/12/06/push-notification-spying/](https://9to5mac.com/2023/12/06/push-notification-spying/) --- ## Post 2 by @anon89321548 — 2023-12-06T17:50:26Z This answered a question I had: _"What can push data reveal?_ _The first important point to make is that, if you are using end-to-end encrypted messaging services like iMessage and WhatsApp, that encryption still protects the messages – even if you’ve set your iPhone to preview the content."_ --- ## Post 3 by @anon5868128 — 2023-12-06T17:57:29Z That’s good to know. Is it any different for iMessage texts synced to iCloud? ~~They’re not encrypted in the cloud~~ , so are they encrypted when they hit Apple’s Push Notification Service…? Also curious if Apple will change anything about how push notifications are processed going forward. I don’t think most people considered this as a potential security vulnerability. --- ## Post 4 by @jonah — 2023-12-06T18:16:04Z They are encrypted both in the cloud and within APNS yes. --- ## Post 5 by @anon5868128 — 2023-12-06T18:23:41Z Apple has a decryption key for iCloud backups (which can include iMessage), no? > **[iCloud data security overview - Apple Support](https://support.apple.com/en-us/102651)** > > iCloud uses strong security methods, employs strict policies to protect your information, and leads the industry in using privacy-preserving security technologies like end-to-end encryption for your data. --- ## Post 6 by @Voln — 2023-12-06T19:02:36Z Apple having a key is very different from the data not being encrypted. Further, the link you shared specifies that one can enable Advanced Data Protection to become the sole key bearer for iCloud backups. --- ## Post 7 by @anon90831229 — 2023-12-06T20:48:39Z We need something like [https://unifiedpush.org/](https://unifiedpush.org/) but as first-party citizen of all OSes, mobile and desktop (though not that important for desktop, I see no reason why not make it an option). Let people use Apple or Google Push Service if they want. But let them choose another provider, or self-hosting etc. Centralization for push notifications obviously gives some benefits when it comes to energy usage, but it doesnt have to mean we all need to rely on the same two centralized systems. --- ## Post 8 by @dngray — 2023-12-07T12:47:44Z This isn’t exactly an unknown thing by developers. There’s a reason why Signal doesn’t put the message in the [`sendNewMessageNotification`](https://github.com/signalapp/Signal-Server/blob/45848e7bfe1813924c677ab06402fb921913d8b2/service/src/main/java/org/whispersystems/textsecuregcm/push/MessageSender.java#L89). With Matrix [it’s optional](https://github.com/matrix-org/synapse/issues/1540#issuecomment-948710624) (for rooms that are not encrypted) as someone self hosting might also have their own [Unified Push](https://unifiedpush.org/) server that they trust. --- ## Post 9 by @ph00lt0 — 2023-12-07T13:35:58Z I am surprised that everyone seems surprised about this. This stuff we have known for a while now. Metadata is a thing, and the feds do force the tech industry to share it. Nothing new to see here imho. --- ## Post 10 by @jonah — 2023-12-07T15:45:39Z The news is the gag order, not the surveillance. --- ## Post 11 by @ph00lt0 — 2023-12-07T15:58:54Z Yeah but that’s nothing new either or is it? I am failing to see this as of yet. All of these programs were/are secret. --- ## Post 12 by @anon89321548 — 2023-12-09T06:40:38Z Let’s say you use an app where the message contents are encrypted on the push server. For example, Signals ability to disable the message contents. (I know their implementation is different this is just an example.) This means that the contents are inaccessible, but the foreign party still has the encrypted form of the message. If you set the app to “Name only” vs “Name and contents” does that mean the push notification would only contain the encrypted information of the name and not the message contents as well? In this case, the user intending to remove the encrypted form of the message entirely as it is possible for the encrypted message to be compromised at some point in the future. --- ## Post 13 by @jonah — 2023-12-09T19:05:30Z It _should_. It depends on how the app implemented push notifications on their server end. --- ## Post 14 by @anon28734771 — 2023-12-11T08:17:19Z One more reason to use F-Droid. > You need proprietary Google Play Services on your phone for notifications to work, apps need to have proprietary Google code in them for this to work, and this is a proprietary service. > You’re purely depending on Google and their proprietary services for basic things such as receiving notifications, which is pure madness. They also probably see from what apps you’re receiving notifications and when you receive them. > Good luck receiving notifications with this without FCM: [Release Element Android v1.6.5 · vector-im/element-android · GitHub](https://github.com/vector-im/element-android/releases/tag/v1.6.5) > There are quite some apps that only provide alternatives to FCM, Google Apps, etc in their F-Droid build. > [@F-Droid (FOSS Android App Store)](https://discuss.privacyguides.net/t/f-droid-foss-android-app-store/13650/20): > > Everybody has different preferences, opinions, threat models, etc. I don’t see a point in arguing about this anymore. A higher target SDK does minimize the risk; that’s true. But again, I trust the F-Droid team, and I could give root access to their clients, just like I fully trust GrapheneOS and use their OS. At this point, it is irrelevant to me if their client has a lower target SDK. When I was writing that post I predicted that someone might mistake alphabet boys (NSA, FBI, etc) with A… Also, this is why FCM and APNS are an issue even if the contents are encrypted: > **[flawedworld | @flawedworld@infosec.exchange (@flawedworlddev)](https://nitter.net/flawedworlddev/status/1733971903553970241#m)** > > If your app relies on FCM/APNS for push notifications, then just encrypting the content is not enough. Your backend server needs to be serving decoy notification traffic to throw off threat actors observing traffic on the wire, and then using... --- ## Post 15 by @Valynor — 2023-12-13T03:04:40Z How Signal handles push notifications: > **[Meredith Whittaker (@Mer__edith@mastodon.world)](https://mastodon.world/@Mer__edith/111563865413484025)** > > PSA: We've received questions about push notifications. First: push notifications for Signal NEVER contain sensitive unencrypted data & do not reveal the contents of any Signal messages or calls–not to Apple, not to Google, not to anyone but you... --- ## Post 16 by @Paced0594 — 2023-12-13T14:34:02Z Thank you for sharing! I was wondering about this. --- ## Post 17 by @anon32876053 — 2023-12-17T09:46:10Z Apple is not alone with the spying on push either, Google does the same. --- ## Post 18 by @anon89321548 — 2023-12-17T09:52:54Z Notably, IMO, is Google came out and said they always required a court order where Apple was not. Considering Apple’s big talk on privacy it’s a pretty huge miss they were not requiring a court order to begin with. --- ## Post 19 by @anon32876053 — 2023-12-17T13:14:27Z So true! However, did Google statement hold true about the court order before or after what ever it had gleam from users. These companies talk the talk, but what do they do ? Well, that’s anybody guess. --- ## Post 20 by @anon89321548 — 2023-12-17T13:19:19Z > Google spokesman Matt Bryant said the company has always “required a court order” to compel disclosure of data associated with push notifications. Regardless of the truthfulness: Apple did not require a court order and IMO, they should have. --- ## Post 21 by @anon90831229 — 2023-12-17T16:15:26Z > [@anon28734771](#): > > Also, this is why FCM and APNS are an issue even if the contents are encrypted: > > [https://nitter.net/flawedworlddev/status/1733971903553970241#m](https://nitter.net/flawedworlddev/status/1733971903553970241#m) Yeah! Not sure why nobody is talking about this (or at least not enough people). I see so many people being mainly concerned with the content of push notifications (_ehem Meredith Whittaker_) but not addressing the far more interesting concerns. Everything from that Xitter thread, but also I would be interested if services make any effort to make it harder to correlate a push token from Apple/Google with the actual user account. I could imagine some solutions for that, not really trivial though. Because even if Signal does not store a lot of data that they can give out about user accounts, if LE can trivially link that to an Apple ID, you might get into trouble through that. Not what I expect from a privacy-conscious communications app. Well, personally of course I’m on google-free GOS anyway and nothing touches FCM or Apple’s Push servers, but still I also need protection for the people I talk to. --- ## Post 22 by @anon89321548 — 2023-12-17T17:47:39Z If you disable push notification entirely for an app does this mitigate in some way? I’m thinking something like proton mail where I don’t really need an instant notification of a new message. --- ## Post 23 by @anon90831229 — 2023-12-17T20:22:56Z > [@anon89321548](#): > > If you disable push notification entirely for an app does this mitigate in some way? Another very important question that app devs are not really answering. Hell, even for instant messanging, it would be good (and important!) to at least have the option. Why is there no Signal FAQ entry by now on their site that states this very clearly: Is anything alleviated by disabling push notifications or do Signal servers still send crucial metadata out to Apple/Google servers? The notifications ultimately not reaching the device (or just not being displayed, even that is not being explained) means nothing in this regard. It should be transparent what information is given from the service provider to Apple/Google and in what exact circumstances. This is not Farmville, but these are alledgedly very security and privacy-focused communication solutions. --- ## Post 24 by @jerm — 2024-01-21T09:36:07Z > [@Apple reveals 'push notification spying' by foreign governments](https://discuss.privacyguides.net/t/apple-reveals-push-notification-spying-by-foreign-governments/15440): > > [https://9to5mac.com/2023/12/06/push-notification-spying/](https://9to5mac.com/2023/12/06/push-notification-spying/) Adding how each messenger handles push notifications and how much metadata the leak would be great addition! Examples: > **[Meredith Whittaker (@Mer__edith@mastodon.world)](https://mastodon.world/@Mer__edith/111563865413484025)** > > PSA: We've received questions about push notifications. First: push notifications for Signal NEVER contain sensitive unencrypted data & do not reveal the contents of any Signal messages or calls–not to Apple, not to Google, not to anyone but you... > **[Frequently Asked Questions - Session Private Messenger](https://getsession.org/faq#push-notifications)** > > Session's FAQ. Find answers to some of the most frequently asked questions about Session — info on the team, the tech, and the technicalities. > And for Briar and Element. --- ## Post 25 by @dngray — 2024-01-21T09:56:13Z This was already discussed in [Apple reveals 'push notification spying' by foreign governments](https://discuss.privacyguides.net/t/apple-reveals-push-notification-spying-by-foreign-governments/15440/) and not much of a concern for the products we list. --- ## Post 26 by @ph00lt0 — 2024-02-10T12:24:29Z Well true, actually might be a good idea to reconfirm this on the website as many people have apparently heard about this story. It could be a good criteria for a messenger. --- ## Post 27 by @ph00lt0 — 2024-03-04T12:25:01Z Also relevant here : [https://www.washingtonpost.com/technology/2024/02/29/push-notification-surveillance-fbi/](https://www.washingtonpost.com/technology/2024/02/29/push-notification-surveillance-fbi/) --- ## Post 28 by @jerm — 2024-03-04T13:29:28Z no paywalled version [https://archive.is/USTPM](https://archive.is/USTPM) --- ## Post 29 by @jerm — 2024-03-04T18:14:12Z Interesting, Teleguard messaging app that is made by [https://swisscows.com/](https://swisscows.com/) helped identify the person by providing push token to the investigator > The document says only that the investigator “provided” the token “as received from TeleGuard,” without explaining how. --- ## Post 30 by @jerm — 2024-03-23T16:50:38Z > **[Why tf does Proton Mail need Google Play Services to show notifications? -...](https://discuss.grapheneos.org/d/9968-why-tf-does-proton-mail-need-google-play-services-to-show-notifications/15)** > > GrapheneOS discussion forum > Tuta was way ahead in this department. Proton is already susceptible to Apple/Google Push notification theft, since they also reveal the mail titles as well as the sender mail, which is very insecure tbh --- ## Post 31 by @moonwriting — 2024-03-23T17:50:22Z > [@jerm](#): > > Tuta was way ahead in this department. Proton is already susceptible to Apple/Google Push notification theft, since they also reveal the mail titles as well as the sender mail, which is very insecure tbh This claim is false and has already been refuted in the thread. Proton encrypts notifications on both [iOS](https://proton.me/blog/ios-security-model) and [Android](https://proton.me/blog/android-client-security-model) so that neither Apple, Google nor any government can spy on them. --- ## Post 32 by @Tech-Trooper — 2024-03-24T23:06:31Z Even though Signal supports notifications outside Google Services, it also encrypts them like Proton. No difference, afaik. It depends on the implementation. --- ## Post 33 by @anon73886004 — 2024-03-30T22:10:46Z Even encrypted it can be preferable to avoid routing the encrypted message contents (by way of the notification) through extra servers. The implementation of this is not always made clear. I believe on iOS just simply disabling show previews does not mean the notification preview, and thus message content preview, are not still included in the encrypted contents of the notification. > **[Harvest now, decrypt later](https://en.wikipedia.org/wiki/Harvest_now%2C_decrypt_later)** > > Harvest now, decrypt later[a] is a surveillance strategy that relies on the acquisition and long-term storage of currently unreadable encrypted data awaiting possible breakthroughs in decryption technology that would render it readable in the future – a hypothetical date referred to as Y2Q (a reference to Y2K) or Q-Day. > The most common concern is the prospect of developments in quantum computing which would allow current strong encryption algorithms to be broken at some time in the future, maki... --- ## Post 34 by @jerm — 2024-07-17T10:48:19Z > [@How Secure Messaging Apps Leak Sensitive Data to Push Notification Services](https://discuss.privacyguides.net/t/how-secure-messaging-apps-leak-sensitive-data-to-push-notification-services/19508): > > [https://arxiv.org/pdf/2407.10589](https://arxiv.org/pdf/2407.10589) Like most modern software, secure messaging apps rely on thirdparty components to implement important app functionality. Although this practice reduces engineering costs, it also introduces the risk of inadvertent privacy breaches due to misconfiguration errors or incomplete documentation. Our research investigated secure messaging apps’ usage of Google’s Firebase Cloud Messaging (FCM) service to send push notifications to Android devices. We analyzed 21 popula…