> They say that anyone who controls WhatsApp's servers could effortlessly insert new people into an otherwise private group, even without the permission of the administrator who ostensibly controls access to that conversation.
Given that WhatsApp isn't open source and so on, controling the WhatsApp server, or controling some part of its code, or controling the signing keys, all can compromise the privacy and encryption.
You still trust the server, unless the encryption is done with code that wasn't delivered from the server. E2E prevents your content from being stolen in a data breach or from being accessed if the server was fine when you sent a message but compromised later.
Good point. The (variously named) security code should allow you to withdraw even that trust (assuming you verify the security code and the binary on your client...), right. Or does it? If the server knows the secret, it can invisibly MITM you, right?
I have. It is a pain in the ass but certainly doable.
I am experienced with iOS, but honestly it is a big app so those who are familiar with Android could do this better, as I believe they actually have decompilers versus needing to read the compiled ARM code.
But that in itself is also a problem, because people outside the security community does not like Signal.
I've tried so many times to convert people, but every non-security person prefer LINE, WhatsApp and Telegram. The reasons are simple; they are fun to use, good user experience and the most important one "my friends use it".
In the past, I've thought of installing Signal, but never did because it identifies you by your phone number, which seems like a totally unjustifiable and idiotic idea. I'd like to keep using my internet-based service even if I switch phones. There is no reason for my phone number to be known to Signal at any point.
Just now, I tried to install Signal and WhatsApp for work communication. Neither is willing to let me do anything unless I grant them permission to read my contacts.
I use WeChat. (Not for security reasons, obviously.) Somehow, despite the fact that WeChat is heavily coupled to your phone, they've realized that it makes more sense to use a WeChat account than a phone number, and that it makes no sense to prohibit users from just adding contacts manually. Adding contacts manually is actually a preferred approach! (In the form "scan QR code".)
Yes, "my friends use it" argument greatly impedes switching to another app, even though it is better. Anyway, slowly, but surely I expand my circle of Signal users. Current version of Signal is actually quite polished, at least not worse than WhatsApp was the last time I used it (before acquisition by Facebook), and I rarely hear complaints about usability anymore from people who have actually tried it.
Yes, in Signal you for example still cannot share one photo to multiple recipients or share multiple photos. I've installed it to all my relatives and they use it anyway, just those simple things would make it a bit better for your average "share pictures of my kids" types of use.
People say things like this about Signal but tend not to acknowledge why Signal is like that. Look at how Signal handles something as basic as user profiles, then compare it to how other applications address the same problems. I'll recommend Wire alongside WhatsApp any day, but keep in mind that Wire's servers apparently have a record of every conversation that has occurred between any two Wire users (not the content, mind you, just the link).
This is why I disagree with Matthew Green, do not think we've totally figured out secure messaging yet and that they're all "so good", and think that if you're serious about privacy --- enough to have strong opinions about WhatsApp vs. Signal, for instance --- that you should use multiple messengers:
- a "tier 1" secure messaging app like Signal that makes all reasonable tradeoffs in favor of security and privacy regardless of the UX cost, used when possible and for sensitive conversations.
- a "tier 2" secure messaging app like WhatsApp or Wire as your "daily messenger".
- "tier 3" messenger applications (including email) that you use mostly to rendezvous to a real messenger application.
In this scheme you can start to understand Signal as not just a decent messenger application with best-in-class security and privacy, but also as a laboratory for future privacy enhancements to messaging.
Ok, just those problems I have outlined seem to me like an UI hurdle. What's the problem making a simple copy of the message dispatched to the contacts I choose? As if I would do it manually?
but messaging is only a small part of whatsapp success. Whatsapp is a social network site. where you add friends by the bucket, even if you dont't plan on messaging them, and share baby pictures to a circle or family/friends.
And maybe add some "defence in depth" by splitting really sensitive ("tier 0") stuff across multiple channels/apps, such that any one breach doesn't compromise it.
Doesn’t this have the potential to inform a potential adversary as to your habits regarding those levels? “Ah he’s back on “tier 1” so this might be juicy!”
You can share photos 1 to 1 in signal too. If the other person doesn't have signal then you have to set your providers mms server in the settings, but it's not very difficult.
Signal has the perception of being more secure simply because it isn't owned by Facebook. I will never trust a Facebook server with private messages, despite both Signal and WhatsApp having the same security features on paper.
* stopped all other monetization efforts (paid api gateways etc)
just to provide free messaging service to everyone.
I have two explanations:
* either he felt it was a threat to his future messenging monopoly
or
* (and this is already not a secret anymore) they wanted to feed the data into their already huge tracking and ad serving network.
Both of those are good enough reasons for me to leave as I care about healthy competition and my future privacy.
But maybe the biggest reason why is because they lied to me: they promised to be the service that provided a good messaging service in exchange for a modest fee. They were profitable and yet sold out.
"An evil plan" might be a bit of an exaggeration, but I think he might have expected more (read more private data or ad push) from WhatsApp users than he currently can get. Of course, that's just speculation, but as you've originally mentioned, it makes no business sense for Facebook to maintain a free messaging app with no strings attached.
Purely on the fundamentals, you could also start with the published source code, build it, and see whether the output matched the binary you're auditing.
In practice, that's unlikely to work, but it could work.
The problem isn't verifying that the source was used for some particular binary, the problem is that when you read source code, the names of functions and variables will impact your understanding of what they do. If you were to take the source code and remove all comments and randomize every symbol name, then you might be okay.
That's a problem, sure, but it doesn't suggest at all that "real security analysis" means starting with an obfuscated blob and reversing it. It suggests that you're better off doing both kinds of analysis. Variable names can lead your thoughts in certain directions and make it more difficult to see certain execution paths. Obfuscated blobs won't lead your thoughts much of anywhere they weren't already inclined to go -- but they make it much more difficult to see all execution paths.
Good luck not being influenced by comments (although they can be stripped) or by how things are named. I still think you are better off starting with a binary.
Like others, I specifically picked Signal over WhatsApp due to the problem of trusting proprietary apps. I think that the UK Govt Ministers that use WhatsApp to coordinate their actions should switch to whatever GCHQ recommends for givt business – I'm sure that they haven't approved WhatsApp.
All that I remember was that he helped with the implementation and that they have a functioning system that they consider correct.
I think he and everybody else knows the limits of the security model of WhatsApp. Just because they have a working setup, does not mean they are actually using it, or have not changed it without his knowledge.
His assertion was simply that WhatsApp was capable of offering the same service as Signal.
This is from memory, but I don't remember any 'PR tour'.
It doesn't matter about preference. Many security people will tell you "WhatsApp is encrypted" as if that makes it secure. My feeling is that rather than deriding it as the false sense of security that it is, the community prefer to trust its claims and believe that it's secure.
Yes many security will tell you "WhatsApp is encrpyted" and that is true, and it is simply your assertion that this implies that they believe it is secure.
But its not just a false sense of security. Yes, everybody that understands security will have issues with all the problems all of us know that exists. That however does not change that for user there was a real increase in security.
I would rather say it is rather disingenuous for you to claim that it is all false security just because it is not what you would consider perfect.
And that is exactly what I mean by security people giving up. Twenty years ago what I think is perfect was merely the norm. Now WhatsApp is OK and it's just me who doesn't consider it perfect.
I was just at a confrence where new more secure communications was one of the topics and there are many people working on new apps, improved protocols and working on figuring out flaws in existing products.
Everybody there understands the security constraints of Whatsapp, and believe me they can hardly shut up about it.
Exactly. Whatsapp "saying" they use e2e encryption itself is a suspicious thing. I don't know what algorithms they are using. I don't know they're not taking fucking screenshots of the chat messages and piping them to Zuckerberg's personal machine. Talking about encryption in a closed source app in istore or android play (which don't allow you to even verify binaries yourself) is just like discussing how well the infotainment system works on a flight with a burning jet engine.
Given that WhatsApp isn't open source and so on, controling the WhatsApp server, or controling some part of its code, or controling the signing keys, all can compromise the privacy and encryption.
This study shouldn't cause much surprise.