Scientist Predicts a Catastrophe in Crowdsourcing
In an interview with Research Scientist, Manuel Cebrian, from the winning team of the DARPA Red Baloon Callenge & the Tag Challenge, he revealed an ominous prediction. He predicted that crowdsourcing, like other high-profile technologies (nuclear, genetic-engineering, etc.), is due for an event, a bad event, a catastrophe, no less.
The reason that this made me a little nervous is that Manuel is speaking from experience; his team has won two contests involving crowdsourcing strategies to achieve the seemingly impossible. In the Red Balloon Challenge, his team found ten weather balloons placed around the U.S. And, in the most recent challenge, found 3 out of 5 specific individuals hiding out in public in major cities across the world.
Only after Cebrian had seen the success of his method, the same method used in both the balloon & tag challenge, did he really suspect that he had stumbled upon something significant in the manipulation of the science of networks that could, if used maliciously or irresponsibly, result in something terrible.
Here's a break down of the 4 points of concern:
- There's a Tipping Point: All you need to do to make something go viral and reach the crowd is to reach a 'tipping' point # of participants. Along with this goes: it is easier for news sources and famous people to get their content, or call to action, to the tipping point because they are the larger nodes in the system. However, as his team has proven, it is possible to reach this tipping point, getting whatever message you have out there with brute force, in the case of crowdsourcing a clever affiliate program. (I get $500 dollars to find someone, so I'll give you $250 if you help me and you find him, etc.)
- Spam Works! In addition to anyone having the ability to reach a tipping point with brute force, you can also get there by spamming, defaming your opponents, and cheating with bots (automatic computer agents that spam on your behalf). This means, not only could false or harmful information reach tipping point and reach the general public, but it could be done so largely artificially and by very few people, not a great recipe.
- There is Such a Thing as Knowing Too Much: in a surprising analogy, Manuel asked me to recall the last scenes from Batman, The Dark Knight. This is where the masked hero uses all of the information from cell phones to map everything around him, giving him super vision through walls, etc. As a data scientist, Cebrian says this scenario is not completely off the mark. He insists that data like that is out there, and he is surprised it has not yet been used to a powerful effect by some villain or rogue. Not only that, if crowdsourcing could similarly organize individuals in a network, they can give you a lot more information than simple location. Someone who had access to a gigantic human network could have similar power.
- It took 10 years to find Osama Bin Laden, but could it have happened in 10 days using crowdsourcing? Both the balloon challenge and tag challenge bring up this interesting, somewhat unsettling point. Will it soon be too easy to find just about anyone whenever we want? To think of it another way: Crowdscanner was a project where it's fun to find actors wearing a particular T-Shirt out in public; it would not seem as innocent if it took the form of say... "FindYourExBoyfrind.com" (Luckily not a real website), etc.
Whichever form it takes, Manuel, though hopeful that it won't, is convinced that something bad will happen in crowdsourcing before not too long, whether it is an exploited government innovation challenge, or just an evil use of network science.
He says that, in his worries, he is reminded of nuclear scientists and AI scientists who, in the course of their research, have become convinced that perhaps we should slow down, and that maybe we are dealing with more power than we can currently handle.
What do you think? Is crowdsourcing becoming too powerful, or too dangerous? Let us know in the comments.