A Dangerous Idea: Continuous Metadata Sousveillance

watchthewatchers9I’ve been thinking about Zimmerman’s Law:
“The human population may not be doubling every eighteen months, but the ability of computers to track us doubles every eighteen months.” “The natural flow of technology tends to move in the direction of making surveillance easier.”

He seems to think legislators need to “do something”. But I think you need to work with Moore’s Law, or be crushed by it. Legislators aren’t going to help; they’re actually who you’re trying to defend yourself from!

My instincts are that eventually, computers will be so powerful, networks so capacious, that basic data will be completely impossible to keep under wraps. In that scenario, anyone trying to hide anything will be detectable with a bit of signal processing. Secret organisations will be plainly visible via the negative space they leave in the general data exhaust.

Unfortunately we’re not there. My guess is that we’re about 20 years away from something like that. In the meantime, there’s this massive disparity between what institutions have access to in terms of data and what we have access to.

That disparity is potentially quite dangerous, particularly if it’s completely asymmetrical, as it threatens to be. If it were even a little more symmetrical, I believe that large, secretive institutions would have far more to worry about than regular people. After all, if a bit of your personal information leaks onto the ‘net, it’s just about always going to be harmless. If a bit of the NSA’s private information leaks, all hell breaks loose and they’re suddenly in existential peril.

What we’ve discovered recently is that the content of communications isn’t all that important. It’s the metadata that let’s you see the general shape of things, the big picture. That’s why the secure email services are shutting down.

You’d think that we’d be able to use metadata in the reverse direction; see into the three letter agencies by analyzing the big data, seeing them in their exhaust, and in the negative space. But we don’t have access to datasets from cell phones, from cloud providers, from interaction with government agencies. We don’t have enough ability to touch the big data.

But we could make our own big data. That’s where Sousveillance comes in.

Sousveillance is “watching from below”, the counterpoint to Surveillance. Up till now, everything I’ve seen people say about Sousveillance has been around Video. But video sousveillance has a lot of problems: video is large (hard to upload in volume), unwieldy (hard to extract information), and recording video is still difficult to do continuously and inconspicuously.

But we’ve just learned that “metadata” is actually what’s useful. It’s not the content of an event, but the time, date, participants, location, devices involved. All that stuff is what you actually want to extract interesting big data signal.

All of us now carry devices capable of metadata sousveillance, right now. They’re our mobile phones, tablets, laptops, and soon to be watches, glass, and other wearables.

On these devices, you can monitor all of your own communications. But you have quite an array of sensors. One of the most interesting and most often forgotten is your network hardware. Your network hardware alone is aware of devices around it. Mac addresses of wifi devices, bluetooth devices.  Services advertising themselves on local networks. NFC devices and tags you interact with. Cell towers and related metadata. And etc.

Take that kind of data, stamp it with location, identity and timestamps, and push it online. All the time.

With the right app or apps, your devices could voluntarily upload streams of metadata to public repositories on the net. Users should be aware and voluntarily participating, but needn’t actually be technical. Just install and go.

With enough people installing such software, the repository we’d get would grow stupendously. And you’d start to see things. Maps of devices inside buildings being picked up by people walking down the street. Clusters of otherwise unrelated mobile devices turning up together in the same places at the same time. Protestors might start mapping devices used by the police, turning up from one place to another. And what else? I’m not sure, but I’m pretty sure there’d be amazing secrets to be uncovered.

Early on it’d likely be fairly dangerous to be involved, because you’d be pretty exposed. You’d be posting your own information freely online, after all. But if the idea spread, it’d start to be safer and more powerful I think.

It’d start forcing secretive institutions to try to obfuscate themselves, or else stop using open protocols. Both paths would really damage those institutions, making them less able to operate in the modern world.

There’d be some pretty amazing technical challenges. Where does this data go? How do you handle, store this massive stream of stuff?

But I think it’s probably doable. And it’s probably necessary, if we’re to push back against Zimmerman’s Law.

A Dangerous Idea: Continuous Metadata Sousveillance