Thursday, July 18, 2013

The Zakkit Privacy Project

I'm trying a little experiment that I'm calling "Zakkit". It's designed to improve people's privacy. Here's why I'm doing it.

I've written a lot lately about the massive amounts of data being collected by the United States government, which includes phone records and vast amounts of internet traffic. I'm not happy about these programs -- they strike me as obviously unethical invasions of privacy, and the arguments to the contrary are ill-informed, disingenuous, and dangerous.

What makes the privacy issue so difficult is that it has a lot of different facets. There's a thorny legal question about whether these data-collection programs are legal; there's a tough constitutional question, too. There's an argument about whether the potential trade-off of privacy and security is worth it (or whether, indeed, there is a trade-off between the two). And there's also a minefield of political issues involving public opinion and the apparent willingness of our political leadership to undertake these efforts out of the public spotlight.

But there's also another problem that's not nearly so difficult -- this is merely a technical problem. Is it technically possible to protect one's own communications against government eavesdropping, or indeed, any other kind of eavesdropping, including by employers, internet service providers, and hackers? The answer to that question is that it certainly is possible to do this. In fact, there's been technology around for a long time that enables people to communicate securely, regardless of the resources that others might spend trying to eavesdrop. The technology that allows for secure communication is free, it's been extensively tested, and there are no known defects.

The trouble is that this technology -- encryption -- is really difficult to use. It takes a lot of understanding of the technical underpinnings in order to allow for real, two-way secure communication. The software that's out there is often too difficult to use, and it requires that all parties to the communication use it correctly. Because it's so hard to use, almost nobody does. Instead, we all communicate over email, which is probably the least secure communication method we've ever invented.

This bugs me because we don't need to wait for the difficult ethical, legal, and political issues to be solved. In principle, we could all communicate securely right now, this very minute. And if our government, employers, schools, and internet service providers wanted to keep on eavesdropping, it wouldn't matter at all. Their efforts would simply fail.

This is why I want to try an experiment. I want to take one form of communication and make it secure. Specifically, I want to let people send documents to each other securely, using state-of-the-art encryption, but wrapped in a service that's as easy to use as email. The basic idea is simple: create a website where anyone can sign up, and send or receive documents to each other by going through a process that's just like sending an email attachment. The difference between this and email, however, will be that the system will deploy state-of-the-art encryption to protect the document against any eavesdroppers. In fact, it can be designed so that even the people with access to the secure servers won't be able to read any of the documents that are sent using this system. People will know that it's secure because it will use open-source encryption software, that's been tested and examined by security experts for years.

The cost of such a service would be minimal. It could be done on a "freemium" model, like Dropbox. Casual users wouldn't have to pay anything. Heavy users, or institutions, could pay a modest monthly fee for "all you can eat" service, maybe around ten bucks a month. Sending a document securely could be totally free, and we wouldn't even require senders to sign up. Storage requirements would be small, because the system would securely wipe out all documents after they've been received. Computational requirements would be modest, too.

At any rate, I've put up a website for the service, and I'm writing and testing the code now (with very promising results). I've immodestly called it "", and I'd love to get any feedback from anyone interested. If you sign up on the site, you can get updates, and be first in line for a free beta. If there's any interest at all, we'll go live on September 1st. If you're interested, why not help out by sharing the site with a few thousand of your closest friends?