VoIP Numbers Stations were Social Experiment

IO ERROR writes "The mysterious phone number stations appearing on Craigslist for the last three months, which resembled their shortwave radio cousins, and which Slashdot reported on in June, were an experiment devised by security researcher Strom Carlson and a group of Los Angeles hackers to determine if encrypted messages could be passed using unwitting third parties to foil traffic analysis by hostile intelligence agencies. Carlson and the hackers presented their findings at DEFCON earlier today and gave away CDs with "Make your own Mein Fraulein station" kits and posted one final number station for people to try to decrypt."

Comments

Not much of an experiment

[ericpi]

experiment ... to determine if encrypted messages could be passed using unwitting third parties They managed to share anonymous information with others using only a site whose purpose is to share anonymous information with others. The fact that the encrypted the info still doesn't make it much of an 'experiment'.

Re:Not much of an experiment

[Anonymous Coward]

How many of them went to their friends and said "check this thing out"? After a few thousand calls or a few million pagehits (HELLO WORLD), I'm sure the person the message was intended for got it, but where I do start looking for that person? Thats a lot of log files to go through and people to extraordinarily rendition before I find the one who knows what the message said. Of course, I'd have to torture each one thouroughly just to make sure they're not lying before I toss their body into a shallow grave and move on to the next. Of course, I'll have to spend some extra time with the men to make sure that they appreciate the homosexual rape and the electrodes and the broom hand... hang on, I've got a call from the President I have to get.

Re:Stenography Encryption

[zacronos]

I think that was the point... As GP said, "the most important thing will not be to make sure that the government can't read what you communicate, but rather have no reason to suspect you're doing anything they don't like".

In other words, you'll (additionally) need to hide your communications, not just encrypt them. If the government doesn't know any of your encrypted traffic exists, or can't attribute it to you, then there would be no case for a visual tail, possibly excepting the "This person seems to have no additional traffic... that's impossible! Must be hiding it somehow. Put a visual tail on 'em." uber Big Brother possibility.

That's the Holy Grail, pretty much.

[Kadin2048]

You've just hit one of the biggest problems facing intelligence today square on the head.

In times past, the real trouble was in the acquisition of information. Now, the problem is on the analysis end: there's just so much information pouring in, nobody can even store it all, much less analyze it to any significant degree. You've got signals from the radio spectrum (broadcast TV and radio, satellite signals, telephone signals), plus all the POTS system voice traffic, plus actual Internet data in its myriad formats; it's really overwhelming.

I don't think there's any pat answer to your question. Obviously the intelligence agencies think that the best solution to the problem is with better analysis software and heuristics programs; stuff that can comb through the haystack and try to find the needle. But of course, those systems are only good at finding stuff, if you have a reasonable idea what you're looking for.

International terrorism, which is the bogeyman today, hasn't been around for long enough that -- in my uninformed opinion, anyway -- we probably know exactly what the "fingerprints" of an upcoming operation look like. We've had a couple of incidents to go on, now, but those are precious few datapoints to base future predictions on, or to use in order to seed systems in the hopes of catching future activity beforehand. It will probably be only in hindsight that we'll know of the next few incidents, and we'll have to use those to program the systems to sort the data.

Obviously, it's a very hard problem, both in the literal layman's sense of the term but also I think in the information-science sense of the term. My personal feeling is that it's such a lucrative problem, both in the public and private-sectors, that we'll get quite good in the future at mining through the rough to find the diamonds; however, it'll always be a cat-and-mouse game with people who want to hide their activities, whatever they are.

To go totally out onto a limb for a moment, my (unjustified) feeling is that eventually, the systems for doing this sort of information-processing will be biological in nature; either using some sort of simulated, self-programming neural networks in silicon, or will actually use neurons that have been plugged in to computer systems (literal 'brains in jars,' perhaps). Assuming we start to see the practical limits of information-processing on silicon, I see biological computing as being the next big step forward in information processing, particularly in the areas requiring a lot of heuristic analysis that don't lend themselves easily to more conventional algorithmic solutions. Data mining seems to be one of the few areas that would have enough possible rewards to justify both the risks and massive investment required, at some point in the future, of research and development.

Re:Stenography Encryption

[Lumpy]

not so.

If you have any brain cells you would make sure that your "visible life" was randomized as much as your invisible life. Then your secret transmissions will be missed as you raised the noise floor so much their detection systems will miss it.

the first way to defeat any detection system is to make it go off all the time and the operatores will start ignoring it.