1
$\begingroup$

Let's say I'm making an end-to-end encrypted messaging app with the goal of putting "privacy first." The contents of messages are encrypted, but in the current model, it is still visible to the server who sends a message to who and when they sent it, as well as the general size of the message. Messages are retained indefinitely.

The model could theoretically be adapted so that after a transaction is complete, there is no unencrypted record of who sent the message (but when it was sent/who received it is still there). Additionally, it could be further adapted to encrypt [the equivalent of Facebook friendships]. These things would still be visible to the servers as the server has to moderate this process somehow, but the information wouldn't be retained.

For the purposes of this question, this app has an equivalent to Messenger groups, with all messages encrypted. These messages could be anonymous, but group membership would still be remembered.

To be clear this will launch initially as English-only and target the US/Europe (with the company based on the US, under US jurisdiction where possible). In the long term this could possibly spread wherever there is internet access and countries don’t block e2e encryption.

So far what I have gathered (from research + comments below) is that implementing this could potentially allow terrorists to hide their association, but on the flip side it could also allow states to pursue people they don't like.

Pros

  • More "privacy."
  • Could be a potential safe haven for people being pursued by their governments.

Cons

  • Could make it harder for Western law enforcement to pursue terrorism.
  • Could mislead people into thinking they have more anonymity than they actually do have.

The verdict

Putting this here since this question was put on hold.

So you guys have raised very good points (all credit to commenters below). I’ve decided to try to make things as anonymous as possible. That way if this app turns out successful it can help people who are targets of their own state. Plus the FBI, etc has other ways to investigate terrorism anyway.

Consider this question answered. Thank you all for opening my eyes.

If anybody has anything to add feel free to comment below.

| improve this question | |
$\endgroup$
  • $\begingroup$ This appears to be a subjective/opinion-based/discussion type of question, which is not a good format for our question-and-answer site. "Is it worth it" depends entirely on the goals of your project, and nobody other than you can make that distinction. Will your application fulfill your goals if it doesn't handle anonymity/anonymous acquaintances, and do you have the resources to implement the functionality? $\endgroup$ – Ella Rose Mar 28 '19 at 14:46
  • $\begingroup$ With no comment on the technical setting—about which you should peruse the anonymity bibliography—note that the general definition of ‘criminal’ is anyone the state doesn't like and can find a rule that they have violated. (Find me six lines from the most honest man, and I will find you something in them by which to hang him.) Is your concern that you aren't helping the Gestapo enough to find the acquaintances of those criminals who harbor Jews? $\endgroup$ – Squeamish Ossifrage Mar 28 '19 at 14:52
  • $\begingroup$ @EllaRose I’ve edited the end of my question to address this. To be clear I do have a theoretical way to implement this (revisions are okay since the app hasn’t launched yet, hence I want to decide these things before it’s too late). $\endgroup$ – Potassium Ion Mar 28 '19 at 14:52
  • 2
    $\begingroup$ Nazi Germany was a Western culture, and Hitler was elected by the rules of a Western constitutional democratic republic. Harboring Jews was criminal under Nazi law. It can happen again in the West, and you are asking whether you should preemptively help enable it. The United States already monitors journalists and immigrant advocates to harass them, criminally prosecutes humanitarian aid, kidnaps children to put in concentration camps, and extrajudicially assassinates people over metadata. $\endgroup$ – Squeamish Ossifrage Mar 28 '19 at 15:29
  • 2
    $\begingroup$ Once you've taken a look at the anonymity bibliography, you are welcome to ask specific technical questions about how to conceal metadata using cryptography. It turns out to be rather difficult, which is why most popular messengers don't do it very well, but that's not for lack of trying, as you can see in the anonbib. $\endgroup$ – Squeamish Ossifrage Mar 28 '19 at 15:44