You know what the best way is to see who cooperates with law enforcement and to what level? Court documents!
I've been trawling court documents for the past few months (I'm writing a blog article on this) and I'm yet to find iMessage being used in court (unless the access to the conversation was given by one of the parties). iMessage really does seem secure from legal system point of view.
For a comparison, I've found dozens of court documents from Google, Facebook, Microsoft handing over chat logs from Hangouts, Whatsapp/FBMessenger, Skype.
Anyway, if you want to see the level of cooperation and want to double-check privacy policies, actual court documents are the best way.
Why? They've been using parallel construction to avoid violating NDAs and giving up their capabilities for years. http://www.wired.com/2014/03/harris-stingray-nda/
1) Police don't care that APPLE doesn't want to reveal its methods, and 2) Unlike the Stingray, it wouldn't be illegal for them to use iMessage evidence in court if Apple provided it to them (especially with the use of a warrant!)
So is privacy a good reason to use FaceTime Audio over say, a standard voice call?
How iMessage sends and receives messages
Users start a new iMessage conversation by entering an address or name. If they enter a phone number or email address, the device contacts the IDS to retrieve the public keys and APNs addresses for all of the devices associated with the addressee. If the user enters a name, the device first utilizes the user’s Contacts app to gather the phone numbers and email addresses associated with that name, then gets the public keys and APNs addresses from the IDS.
The user’s outgoing message is individually encrypted for each of the receiver’s devices. The public RSA encryption keys of the receiving devices are retrieved from IDS. For each receiving device, the sending device generates a random 128-bit key and encrypts the message with it using AES in CTR mode. This per-message AES key is encrypted using RSA-OAEP to the public key of the receiving device. The combination of the encrypted message text and the encrypted message key is then hashed with SHA-1, and the hash is signed with ECDSA using the sending device’s private signing key. The resulting messages, one for each receiving device, consist of the encrypted message text, the encrypted message key, and the sender’s digital signature. They are then dispatched to the APNs for delivery. Metadata, such as the timestamp and APNs routing information, is not encrypted. Communication with APNs is encrypted using a forward-secret TLS channel.
Interesting note: "P. FaceTime FaceTime communications are end-to-end encrypted and Apple has no way to decrypt FaceTime data when it is in transit between devices. Apple cannot intercept FaceTime communications. Apple has FaceTime call invitation logs when a FaceTime call invitation is initiated. These logs do not indicate that any communication between users actually took place. Apple has no information as to whether the FaceTime call was successfully established or duration of a FaceTime call. FaceTime call invitation logs are retained up to 30 days. FaceTime call invitation logs are available only following receipt of a legally valid request"
iMessage is not mentioned. Does this mean they are capable of intercepting iMessage?
edit: in the FAQ it says "Can Apple intercept users’ communications pursuant to a Wiretap Order? Apple can intercept users’ email communications, upon receipt of a valid Wiretap Order. Apple cannot intercept users’ iMessage or FaceTime communications as these communications are end-to-end encrypted."