Freedom of Expression, Privacy, and Anonymity
I'm resuming this blog after a one-year break. When reading back the first 24 episodes, which were written "pre-Snowden", it struck me how much the topic of government surveillance was already present in our conversations about internet freedom. But even more so now that we live in a "post-Snowden" world, it seems to be a default assumption that internet freedom projects are about protection from government spying.
Even when you explicitly say your aim is to break the web 2.0 monopolies, many people assume that the ultimate goal of this is to get everyday communication away from the reach of NSA spying. This is also confirmed as Francis summarizes common motivations for redecentralization with privacy as the most important one, before resilience, competition and fun.
The first pre-requisites for freedom of expression are access and know-how: you need access to a device through which you can send your message towards its destination, and you need to know how to use this device to do this. When a government blocks Twitter, it becomes harder for its citizens to express their political views, because they may not be aware of other channels and how to use them. A big task for the hacker community is to provide access and know-how to citizens whose freedom of expression is under attack.
The third pre-requisite is freedom from prosecution after expressing your opinion. This is where anonymity can play a role, and contribute to freedom of expression. When you express your political views under an ad-hoc identity or a pseudonym, successfully erase all traces back to you as a physical person, and for the rest of your life keep it a secret that it was you who published this opinion under this pseudonym, then the politicians who want to shut you up may be unable to find you. Alternatively, you can move your life outside the jurisdiction of these politicians, as Edward Snowden did by moving from Hawaii to Moscow Airport.
Another reason to want privacy is that eavesdropping is immoral and annoying. We want to be able to have a conversation with the audience of our choice, without having to think about who else is able to follow what is being said. Keeping the content of a conversation private does not require any expensive equipment. As an engineering problem, it is solved by for instance PGP. In practice, however, it requires quite a lot of extra effort and know-how, so that hardly anyone bothers.
Ad-hoc Identities vs. Pseudonyms
Hiding the fact itself that you are having a conversation with someone, is quite a bit harder. For your pre-existing contacts and real-world contacts, you could create a set of one-time-use identities, and give these out inside encrypted conversations. You can then shut down your public identity, and only communicate through these unrecognizable identities. As long as you manage to hide the last mile where you retrieve the incoming messages onto a device at your physical location, an outsider will have a hard time finding out when you communicated with which of your friends.
There are also situations when you want to publish a message to an open-ended audience, without exposing your real-world identity. For this, you can use a pseudonym. The Bitcoin software was published this way, for instance, and it is to this day unknown who is the real person or group of persons behind the Satoshi Nakamoto pseudonym.
It's important not to use a pseudonym where what you need is ad-hoc identities. Many people use an online nickname which is unrelated to their birthname. In some situations, this will give some limited anonymity in interactions with other online identities, but in the end it's practically impossible to have normal everyday interactions under such a nickname without the link to your real-world identity becoming known.
In both real-world and online conversations, someone is bound to use your pseudonym and your real name in a careless combination and expose their relation sooner or later. Examples are Bruce Wayne and Jonathan Gilette.
Tor, freenet, and i2p
An important tool for anonymously accessing the internet is Tor. It mixes up your internet traffic by adding a few extra random hops to the route of each request. It is popular with political activists who need to worry about who is watching their actions. Do make sure though that your device is not backdoored, and that you don't reveal your identity by logging in to any of your email or other accounts when using Tor. Tor can anonymize your client traffic, but it is also possible to expose a server via Tor, such that other people can access your server without finding out where this server is physically located.
When publishing important content anonymously, an alternative to running your own server as a Tor hidden service is to publish it on Freenet, or through Wikileaks or a journalist you trust. It may seem paradoxical that in such a case, encrypting and signing your messages with your PGP identity can form an important part of establishing the trust between you and the journalist, so you end up using verifiable identities to ultimately achieve anonymity.
In practice, it is hard to keep multiple identities separate, and not many people use anonymity to protect their privacy on a day-to-day basis. A small group of people use PGP to hide the content of their personal communication from third-party eavesdropping, but this involves publishing your PGP identity, and unless you use ad-hoc identities, this probably actually hurts rather than helps your anonymity. Using a pseudonymous PGP identity may work to some extent, but in practice only one leak will be enough to link all communication under that pseudonym to your other pseudonyms and your real-world identity as a physical person.
So in practice, I think we can achieve two things with online privacy and anonymity: first of all, freedom of speech for important messages, where it matters enough to go through extra trouble. This can be achieved using off-the-shelf technology. The main difficulty here is not technical, but human: you need a good poker face when people keep asking you "Are you [Batman/Nakamoto/...]?".
The other thing we can achieve is an internet where conversations stay between the participants unless these participants choose to leak what was said. For the content of the conversation, this is already achieved by using an encrypted protocol like for instance WebRTC. I am not aware of any software application or protocol that automatically generates and uses ad-hoc identities between contacts, but you could do this manually: Whenever you send an email, use a randomly generated from and reply-to address which you will use only once, and include your signature inside the encrypted payload, so that only the intended recipient of your message will discover that this message was actually from you.
Next: Decentralized reputation systems