An official FBI document dated January 2021, obtained by the American association “Property of People” through the Freedom of Information Act.
This document summarizes the possibilities for legal access to data from nine instant messaging services: iMessage, Line, Signal, Telegram, Threema, Viber, WeChat, WhatsApp and Wickr. For each software, different judicial methods are explored, such as subpoena, search warrant, active collection of communications metadata (“Pen Register”) or connection data retention law (“18 USC§2703”). Here, in essence, is the information the FBI says it can retrieve:
-
Apple iMessage: basic subscriber data; in the case of an iPhone user, investigators may be able to get their hands on message content if the user uses iCloud to synchronize iMessage messages or to back up data on their phone.
-
Line: account data (image, username, e-mail address, phone number, Line ID, creation date, usage data, etc.); if the user has not activated end-to-end encryption, investigators can retrieve the texts of exchanges over a seven-day period, but not other data (audio, video, images, location).
-
Signal: date and time of account creation and date of last connection.
-
Telegram: IP address and phone number for investigations into confirmed terrorists, otherwise nothing.
-
Threema: cryptographic fingerprint of phone number and e-mail address, push service tokens if used, public key, account creation date, last connection date.
-
Viber: account data and IP address used to create the account; investigators can also access message history (date, time, source, destination).
-
WeChat: basic data such as name, phone number, e-mail and IP address, but only for non-Chinese users.
-
WhatsApp: the targeted person’s basic data, address book and contacts who have the targeted person in their address book; it is possible to collect message metadata in real time (“Pen Register”); message content can be retrieved via iCloud backups.
-
Wickr: Date and time of account creation, types of terminal on which the application is installed, date of last connection, number of messages exchanged, external identifiers associated with the account (e-mail addresses, telephone numbers), avatar image, data linked to adding or deleting.
TL;DR Signal is the messaging system that provides the least information to investigators.
sms through signal was not encrypted, how would that even work? how would the signal app even know your contacts were using an app that supports encryption?
Signal accounts are based on your phone number, so if you’re messaging a number that has an account with Signal, the app could see that and would send the message through Signal’s protocols to that Signal account instead of with the SMS protocol which is not encrypted.
You’re correct I should have better worded my point: Signal used to be a single all that someone could install that could handle sending out their regular unencrypted SMS messages and Signal encrypted messages.
Signal also did exactly what you’ve described - auto-enabled encryption when it detected another signal user by phone number.
The net result was more people using encrypted messaging.
I believe once upon a time, back when Signal was called TextSecure, it actually supported this. I don’t remember how it knew the remote party was on TextSecure, I think it may have just let you send anyone encrypted messages, regardless of whether they could read them but I could be wrong. It’s been a long time.