Jonathan Zittrain: I’m Jonathan Zittrain, I’m a co-founder of the Berkman Center for Internet and Society at Harvard Law School, where I’m also a professor of law.
Topic: Evaluating the Ubiquitous Human Computer
Jonathan Zittrain: Well, I think a lot of computer scientists have been thinking about what they call ubiquitous computing, the idea that with really good bandwidth and wireless at that, and fast processors, there’ll be computers everywhere, more than just in your lap on a laptop or in a Hallmark card when you open it and it sings you happy birthday.
I’m interested in a counterpart to that, which I think of as ubiquitous human computing. And that is a trend that’s just starting to get ramped up, where we start to think of the human mind as if it were a server. As if it were a commodity that you could throw at a problem and ramp up by just investing a little bit more money in your project. And we see this in three layers, each of which has an example company at work and each example of which is really interesting to me. I like these companies, I like all three of them. But in total, I see reason to, I don’t know, be bemused about what’s happening.
At the top level of this phenomenon, is a company like InnoCentive. InnoCentive was started by a big pharmaceutical concern, to create a marketplace for companies, like itself, that might have engineering or scientific problems to solve that aren’t exactly Nobel Prize worthy, but are more complicated than just something you could send downstairs and to be able to take those problems, describe them, and put a bounty on their solution--$15,000 to the first person somewhere in the world that can create a molecule that has the following characteristics. It smells like this and it’s machine washable. And there’s a race on then among those who tend to troll that site—a lot of unemployed chemical engineers, among others—and somebody comes up with the answer, submits it, it’s validated, collects $15,000, and in exchange, it’s as if he or she ever had the idea. That becomes the property--lock, stock, and barrel—of the company that commissioned it for the money, big money, that it paid for it.
So that’s sort of on the high end of putting problems out for solution, where you’re basically indifferent to who’s solving the problem, what their motives are, anything else. You may not even meet them or correspond with them; you just get a solution for your money that required a human brain in the middle.
Then one notch down, there’s a company like LiveOps. LiveOps, from what I can tell, is extremely successful, and it’s a company that will employ anyone in America, at the moment, who can get through its test. And the test is administered automatically over its website, a number of questions, reading comprehension, kind of like a mini SAT test, and other languages, if you’ve got them, are tested. And then they also test your computer outfit, do you have a decent internet connection, do you know how to use your PC, more or less, and a working headset. And of many, many people who might apply on one end, only a few fall out after hours of the tests on the other end ready to be, I wouldn’t say hired by LiveOps, because they won’t be employees, but in a working relationship with them, like contractors. And then it’s almost like a video game. You’re in the privacy of your own home, maybe you’ve got your kid in the playpen next to you, you turn on your LiveOps account, and where do you want to go today? Well, maybe when you first start, it’s just taking pizza orders, and you click, you’re given a script, an identity, suddenly you’re working for Pizza Hut and then someone calls what they think is Pizza Hut, it gets routed to your living room, you say, “Welcome to Pizza Hut,” you take the order, you type it in, you click send, and boom, it goes back out to the local Pizza Hut from the person calling, and the pizza gets delivered.
This is an interesting way of working, because it gives you enormous freedom, you can be in your own home, you can plug in or out after every single call. You can work as much or as little as you like. But it also is a lot of, if not constraint, surveillance. The calls and interactions are recorded. There are all sorts of metrics applied, so you’ll be told if you’re keeping up on the number of calls per hour that is expected for this range. And then as you succeed at the little tasks, you get onto more and more complicated tasks. LiveOps said that when Hurricane Katrina hit Louisiana, the American Red Cross came up with an 800 number that people could call for information, try to find relatives. Of course, the American Red Cross didn’t have enough people to staff it, and turned to LiveOps. LiveOps immediately, boom, sent out the word and those LiveOps agents who were working at that moment all could transform themselves into American Red Cross representatives.
That’s the middle zone. The zone below that is represented by Amazon’s Mechanical Turk. This is a service whereby for maybe a penny, or a nickel, tasks that are repetitive, often boring, but that still require a human brain of some kind in order to complete, are put out for bid. And anyone with a Mechanical Turk account can perform the tasks.
So you might be asked to look at a picture and identify what you see in the picture and type in the keywords of what you see. You might be told to go to a website and interact with it in a certain way. Go to this blog and leave a comment with what you think. Go to this Amazon.com review site and leave a positive review, must be five out of five stars. That actually happened, was discovered, and the company behind it roundly denounced for astro-turfing.
And when you put these three things together, I think what you start to see is a way of harnessing mental energy in action that we just didn’t have before. That can be very empowering, but it also makes me think about how easily one could put out the call to call your member of congress, tell them how much you hate or like healthcare reform. And that call would be indistinguishable from that of any regular constituent, because you are a regular constituent. You’ve just been primed by Mechanical Turk to do it, you then report on your call, maybe even send a transcript or a recording along and you get paid your nickel or your quarter or whatever else. And that starts to mean that any mass movement becomes suspect, because we have no idea how it was motivated or who was asked. Now, maybe word could get out that that was going on, but, as it happens more and more, it’s not clear how scandalous it would be, it’s just another set of tools used, just like we expect today that there will be television commercials aired or testimonials from people who have been prescreened on one issue or another.
So as we enter a world in which I imagine a subway car, which everybody in the car is basically traveling in silence, staring at a screen or talking quietly into something, and some of them are engineers, hoping to earn their $15,000, some of them are LiveOps people checking in with their white collar skills, and others are latter day mental click worker sweat shoppers doing one nickel task at a time. I look at that subway car and I see a very different configuration than the one we have today, even if most people are silent, listening to their iPods. So it’s a phenomenon worth keeping an eye on.
Question: Can Facebook be Trusted?
Jonathan Zittrain: Well, when I think about privacy on social media sites, there’s kind of the usual suspect problems, which doesn’t make them any less important or severe, it’s just we kind of know their shape and we kind of know how we’re going to solve them. Those are problems that hinge around whether the site itself, Facebook say, can be entrusted with the information we give it, under what circumstances will it release it to third parties, things like that. And there, I think, standard privacy laws and forces are already at work trying to keep sites like Facebook from going too far and that’s standard stuff about opt-in versus opt-out. There’s a beacon program that Facebook introduced now called, I think, social ads. And it’s working its way along. In fact, I think it’s under special scrutiny, every day, it seems there’s a new Facebook group called, you know, “One Million Strong Against Facebook Invading My Privacy,” so I’m not as worried about that.
But that’s actually a couple of other issues of interest. One is in the name of privacy, and I think in a way that is actually privacy protective, when we see third party developers writing applications that run on Facebook and that interact with your data, Facebook has a way of being in the middle in an ongoing sense. So when that application wants to pull data from your profile, it’s going through Facebook and Facebook can turn it off later and say, “No, no, you can’t get this data any more because we’ve deemed that you’ve broken the rules in some way or we just no longer want to do business with you.” It’s very different than the free-for-all, say, for an application developer on a personal computer who doesn’t have to deal with, say, Microsoft, even if you’re using Windows in order to get to the data on the computer. That’s between you and the software writer.
So in the name of privacy protection, what we see is, is Facebook able to create barriers and restrictions to the flow of information to third party developers that wouldn’t exist in counterpart systems like PC’s? And then you start to see possibly a trade off between privacy and innovation because the more control a vendor has in the middle, between that relationship, between a software author and somebody who wants to run the software, the less freedom a software author has to do whatever he or she wants.
I also see within a cauldron like Facebook, opportunity for really interesting experimentation about new ways of protecting privacy. I don’t know how much thought is behind it, but it seems to me, highly effective the way that Facebook will let somebody tag a photo with a friend’s name, then others who are friend of that friend can perhaps immediately see the photo, and the friend, in the meantime, has a chance to wander back and un-tag it. At which point the person who owns the photo could re-tag it, and then you un-tag it again. But what you end up with is sort of a good enough for government work system where embarrassing photos might go up and somebody pictured in the photo can’t delete it, but can try to see to it that they’re not identified with it in a way that direct search under their name could find them in the photo. Instead, somebody has to be paging through the entire album and say, “Oh, I recognize that guy,” or see a dead tag link.
Innovations like that, I’d like to see experimented with on the internet at large and just as there is a creative commons to let people express, in the terms of intellectual property, how they’d like their work to be used or re-used, if at all, I think there needs to be a kind of privacy creative commons that lets you express your preferences about how a photo of you should be used or a video or something like that, and then give others an opportunity to see whether they want to respect those preferences. At the moment, there really isn’t a system for that, and that’s how sometimes I think videos can go viral, where people featured in them might well wish it not to be so, but those who are contributing to the viral nature don’t even have a chance to realize it and abide by a request for forbearance.
Question: Are social networking sites here to stay?
Jonathan Zittrain: I think social networking is absolutely here to stay. Now, whether or not the label will Facebook forever, depends in part, I think, on whether Facebook wants to try to be less proprietary, be more central to the operation of defining and stewarding identity online.
Question: Linking the virtual and the physical
Jonathan Zittrain: What’s going to be a person’s anchor point in this new virtual universe and how easily can there be applications that make use of that identity, about our features, that we can control, but how easily can somebody author something new and different for that? And that’s why, to me, at the moment here in 2009, Facebook apps maybe aren’t the biggest thing ever, but that’s where I’d be keeping my eye, to see just how permanent something like Facebook will be. And it’s also why I think something like Twitter has become as, I think for a lot of people, baffling popular as it has, because it is so commoditizable, it’s basically a standard for any one person or non-person, to utter statements to the world in short character bursts. And that’s the kind of thing that can be built into any other number of applications.
So I would look next to see a whole range of applications written that more innately and automatically enhance or describe our identities online as we just go about our daily lives. I mean, right now you see applications like Four Square, Meetro, and Looped, where people can say where they are. You check in at a restaurant and then any of your friends nearby know that you’re at that restaurant and can meet you there. Or you connected up with your social network and now you can find, among the 20 closest friends of your 20 closest friends, here is who else is nearby in this restaurant and you can greet each other and meet without fear of one of you being an axe murderer.
And we’ll see more of those services, I think, and more of these kind of mouse droppings of just as you act in the world, right, it’s just one more shift, so you don’t have to actually check in at the restaurant, but set it to auto check in, and so those with whom you have a certain degree of friendship in your profile, automatically know where you are at all times, as you go from one place to the next. And as more and more products have RF ID’s attached to them, it can be effortless to inform the world that you’re having a bowl of Lucky Charms cereal right now because you were just handling the box.
Now, is that a good thing? Who cares who’s eating Lucky Charms, but it always starts with the trivial and the goofy, and then it turns out to have many more powerful applications. When you realize, for example, that one of the biggest barriers to successful healthcare is whether or not people are taking their medicine, because it often turns out that often they’re sick when they’re taking their medicine and not fully of right mind, you start to see how I see, a bowl of Lucky Charms, who cares? But did they take their pills today? Much easier question to answer in the 21st century than it was in the last.
Question: What is the next threat to privacy?
When I worry about privacy I worry about peer-to-peer invasion of privacy. About the fact that anytime anything of any note happens, there are three arms holding cell phones with cameras in them or video records capturing the event ready to go on the nightly news, if necessary. And I think that does change a lot our sense of what is going on in our neighborhoods. I mean, right now we basically smooth over everything we don’t hear about with an assumption that it was an average day and now all of a sudden, anything not average in the day will be available for us to see and indexed by location, by time, and probably by actor. And at that point, I think, our whole perception of what’s going on in our towns or in other towns, starts to change.
I start to think of taking the ingredients we already have and just mixing them together so that on a base of something like Flickr, what you end up with is an ability, as you’re standing in a spot, to say, “Show me every picture taken at this spot in the past ten years that’s been submitted to something like Flickr.” If the spot happens to be a reproductive health clinic or a flophouse or a church, you can start to see how there can extra sensitivities to that. Or you can just cross reference and say, “Show me every picture in which this person appears through auto image recognition, even in just the background,” so as I’m walking down the street through something like Harvard Square of Washington Square Park, if I’m in the background of anybody’s video or camera, then I will be understood as having been there at that time. And I think that’s both powerful and interesting and also an invasion of privacy of the sort that making you worry about whether Facebook is, you know, telling people what product you just bought somehow pales in comparison.
On Controllable Devices
As we shift away from the personal computer to more defined devices that are themselves in touch with their vendors at all times, iPhones, Kindles, nearly every mobile phone, in fact, these devices are much more controllable, not just then by their vendors, but by anybody who regulates the vendors. And suddenly we can see, as was attempted by China in the Green Damn Youth Escort Project, trying to see that every new PC being sold in China would have filtering software embedded in it that talks to the mother ship and can update itself on what it will filter, or even just what it will do generally, to have devices that are already baked that way, they don’t have to be transformed that way by software, but out of the box, that’s how they behave, I think that is a wild card that could change the initial observation that every day things are a little bit freer. And that’s why I’d be keeping an eye on that.
On Spam and the Stock Market
Jonathan Zittrain: With a colleague named Laura Freider, we had a big pile of spam and wanted to know something useful to do with it, and I was curious at that time, much of the spam was touting stocks. Go run out, buy this stock now, here’s your free newsletter. It’s hard to imagine anybody would go out and buy stock on the basis of a piece of spam they got and I was curious then, why are they still sending it?
So we actually ran a study to see, with the cooperation of Pink Sheets, which tracks many of these penny stocks in volume and price over the course of a day, whether we saw a correlation between the sending of these emails, we ending up accruing a pretty big database of them, and the movement of the stock price. And what we basically found was, yes, there is a correlation and if you buy the stock around the time that the email is sent, bad news, it’s already too late, the price is going to go down after you get it. But if you had had the foresight to buy it beforehand, i.e., you were the spammer, probably a pretty good idea, enough people will get it, either because they believed the stock newsletter, that this is a good buy, or because they believe enough other idiots will believe it that they just want to ride the wave, then you do well.
Now, I’ve anecdotally seen less stock spam lately, so perhaps those orchestrating it have moved on to a different business, but it does show that something like spam actually can have an affect, there’s still people out there reading it and perhaps acting upon it.
Question: Is the internet forcing a new set of international laws?
Jonathan Zittrain: I don’t think the internet requires some new compact, but it’s certainly starting to show its age. I believe it was designed at a time, over 30 years ago, 40 years ago, and among a group of people, that they had what I call a procrastination principle. They wanted, in the words later, of Jimbo Wales, founder of Wikipedia, to release early, release often, don’t make the perfect the enemy of the good. Come out with a network, get it a little bit working, get it a little bit better, deploy it one note at a time, let people sign up as they want to within this set of notes, academic and research institutions, and we’ll solve other problems as they come up. And I think that turns out to be a very powerful strategy and not the kind of strategy you see often in inter-governmental and treaty organizations, when they design something, in their minds, I think they’re often designing it for the next thousand years. They want to get everything just right and any problem that is brought up is a potential problem. Why? What if somebody does something illegal over this network? They feel they have to build in some way of redressing that concern before the first bit even flows across it.
The internet, to my eye, was not built that way and that has made it extremely powerful, it made it out-compete other networks. But to me, it’s the procrastination principle, it’s not the denial principle, and that means that at some point, we may have to reckon with some of the problems that were kicked down the road. And certainly as a experimental network is now used for mission critical tasks, people entrust their financial information through it, that kind of thing, we start to see just how well procrastinating can keep working. Can we do everything at the so-called application layer? Keep the network the same but have you and your bank figure out a way to talk more securely with each other? Maybe, but that’s starting to expect an awful lot of you and maybe there are ways to try to make the network more secure in a number of dimensions.
So that’s why I think the next five to ten years in the so-called internet governance space is going to be very interesting. My instinct tells me that the next set of innovations we’ll see around the network will come from unexpected quarters, not from inter-governmental agreements and treaties that somehow are supposed to change the fundamental fabric of the net by phi at.
Question: What is the state of governmental censorship of the internet?
Jonathan Zittrain: Well, when I think of governmental censorship on the web, and this is a project I’ve worked on as part of the Open Net Initiative, along with the Berkman Center, the University of Toronto, and something called the SecDev Group, they’re kind of two contradictory poles. On one pole, I believe that each day on average it’s a little bit easier to push information from here to there. There’s just that many more devices out there, bandwidth is getting better, it’s just harder to keep a secret out there. So on that pole it says, if you are an authoritarian government, and at least you’re not hermetically sealed the way, say, North Korea is, this is making life harder as far as censorship goes.
On the other hand, we do see regimes and bureaucracies that really didn’t have the internet on radar at all. They had strict censorship in place for television, mass media, newspapers coming in at the borders, but just weren’t thinking much about the internet. They now are. And in places like China, we see very sophisticated, increasingly sophisticated schemes for driving people away from information they might be wanting to get and we see a populace that, by and large, is not necessarily up in arms about it. There may be an active, small, free speech segment, but the typical person on the street, when trying to get somewhere and there’s a network error, might not even know that the government isn’t wanting them to get there, but says, “Okay, fine, I’m going to go somewhere else.” Just as we would in another regime, when the network happens to be slow going to Facebook, it’s like, “Okay, I’ll visit Twitter now,” or whatever it might be.
So it’ll be worth keeping an eye as to which of these poles will predominant. And some of the factors that will be wild cards here include how much governments will turn to surveillance, and not just filtering, so that they use the fact that you went to get to a website that might be sensitive as a reason to keep a file on you or augment the file they have on you. And when you put that together with the ability through ubiquitous human computing, to hire tons of people to do small tasks, like keep an eye on X and tell me what you see, in the physical world counterpart, it would be, “Here’s a section of border fence, watch it for the next 20 minutes and if you see anybody, press the following button.” You could actually see people around the world being enlisted to help spy on other people around the world. Not even necessarily knowing what they’re doing or why.
Social media, of course, is also a wild card here. It turns out that I think something like Twitter is a little harder to filter on average, because there are so many different API’s, different ways in and out of Twitter that just blocking Twitter.com doesn’t do the trick, and that a formerly innocuous site, a blog about woodworking, can suddenly become very sensitive because it embeds automatically a feed from somewhere else, a Twitter feed, and can suddenly start carrying Iran election news.
On the other hand, I think what you see from some governments, is an attempt to start mastering those chaotic social media, pumping disinformation in, so that you can’t trust anything you see there. And that might make it harder to have any idea what’s going on, because of the cacophony that you see.
Question: What keeps you up at night?
Jonathan Zittrain: What keeps me up late at night, in the sense of worry, I guess it’s innovation. It’s funny to be worried about it, because it’s a fair point that wow, look at the innovation we’ve seen over the past, not just 30 years, but over the past two years. But I also see ways in which the kind of truly revolutionary applications, like the world wide web, that have come about, kind of unsuspecting to the incumbents, whether government incumbents or corporate incumbents, like big Telcos or something, those get harder and harder to pull off as the devices and services people use are more controllable by a handful of vendors. It’s hard to try to be a canary in the coal mine here because right now we’re enjoying such innovation. But at the same time, we are truly in the midst of a sea change in how controllable the technology we use day by day is, and it is getting more and more controllable by a distinct group of entities, who may have our best interests in mind, at least at consumers right now, but they can change their minds or be regulated, forced to change their minds later.
So that keeps me up at night, and when I think of being kept up at night in a positive sense, I actually think about the possibilities behind so-called ad hoc mesh networking, the ability for these ubiquitous computers, little processors we have everywhere, whether a birthday card or a handheld phone, to talk to each other and like a bucket brigade, get bits moving from one place to another with no internet service provider having to be formally involved in many of the steps. The kinks haven’t been worked out on that yet, but if we had such a thing and could deploy such a thing, because there are enough reprogrammable devices in our hands, that you buy it to be a word processor on day one, but on day five, suddenly it’s got mesh networking, that means that if Hurricane Katrina hits and takes out most networks and power, or if the Chinese government decides to clamp down and try to cut off the internet, neither gets much in the way; because as long as you are near other people, bits are shuffling from one person to the next like hands across America out to drop points and exchange points. So that’s one of the most interesting pieces of basic research that I think is worth keeping an eye on that can transform this constant battle between cat and mouse and how regulable the internet can be.
Recorded on August 18, 2009