 My name is Steve Oliveira. I'm the Dean for Development and Alumni Relations, and it's my great pleasure to ask Dean Minow to come to the podium to introduce our keynote speaker. Thank you. What are you doing? Hey! No, no, we've got to do that one again. Hi, everybody, how are you doing? Hi! All right. Excellent, excellent. So I have met the future. I have met the future. And his name is Jonathan Zittrain. Professor Zittrain's research includes studies of digital property, privacy, speech, the role played by middle people in internet architecture. He has interest in creative, useful, and unobtrusive ways to deploy technology in teaching. He creates platforms where people explore and invent new ways to use digital tools for teaching and learning. For example, in his H2O project, he devises a system that allows people to interact with one another, gather around specific details like a new piece of legislation, and share content across the world instantaneously. In his own teaching, Jonathan Zittrain makes it possible for students to participate in class orally, but also via the internet. He and a team have worked through the creation of OpenNet Initiative to monitor internet censorship by national governments. He co-founded Chilling Effects, which monitors cease and desist letters, and alters actually how internet entities operate. He and colleague John Paul Free founded something called Stop Badware, which functions as a clearinghouse for a proliferation of something called malware, which he will explain to you. One of the goals is to preempt the stifling of the internet. He published an article called Spamworks, which documents the manipulation of stock prices via spam email. Jonathan Zittrain is a professor here at the Harvard Law School. He's also a professor of computer science at the Harvard School of Engineering and Applied Sciences. He's faculty co-director of the Harvard Berkman Center for Internet and Society. Previously, he was professor of internet governance and regulation at Oxford, University of Oxford. His landmark book, The Future of the Internet and How to Stop It, has been described as the most compelling book ever written on why a transformative technology's trajectory threatens to stifle its own great promise for society. And there, he provides a very compelling analysis that studies how our movement from machines that limit their use to a world of computers and internet that opens up every possibility, and then back again to a world of lockdown devices with predetermined apps allowed only by the owner of the creator. This is something that he will explain to you. If you don't understand it yet, you will by the time lunch is over. Jonathan Zittrain defines and shapes the intersection between this whole new digital revolution and what law offers as a framework and as a set of opportunities. But you know what? You'll be reassured, because Jonathan Zittrain can talk like an old-fashioned guy, too. He's wise and he's funny. But I assure you, he is the future. Just listen. Thank you so much, Martha, and good afternoon. I'm particularly pleased to be here since I'm sort of an adopted member of the class of 1995. I started out in 94. Did a program over at the Kennedy School. When I came back, found that I knew almost no one. So I was very pleased that the class of 1995 was so welcoming as I was sort of the wayward person lost between Langdell and Austin. Now, as far as representing the future goes, one opportunity when we have such a vertical representation of so many different marked milestone class years is just lots of eras represented within this room. And today, I want to share some research that I'm doing that I confess this is one of the occupational hazards of studying cyber law and internet law. Scares me a little bit and makes me feel old. But I hope to share that feeling as well. And with luck, the class of 2005 will walk out feeling old as well. In 2009, a graduate student at NYU named Casey Kinzer did a rather unusual experiment. She built a little cardboard robot pictured here, drew an all-important smiley face on it, went to Radio Shack, and got a little motor that propelled it forward at a constant but slow rate, put a flag on the top of it that indicated where the robot wanted to go, and then set them loose on the streets of Manhattan. Now, we just recently saw a lunch pail have Times Square closed for an entire afternoon. And oddly here, the reaction was quite different. People actually noticed the robot and intervened to help it get to where it wants to go. So this is one robot's path through Washington Square Park where over 40 people helped it get from the upper right corner to the bottom left corner. Now, like most sociology experiments, my first reaction was, that's fascinating. My second reaction was, I'm not sure what to make of it. But it resembles a little bit internet routing, the way that packets get from one place to another more like a bucket brigade than FedEx. And that's kind of interesting to me. And it's also interesting the ways in which strangers can be enlisted to contribute to a project in very unusual and unexpected ways. Now, that's something for which there have been a number of books written about this. These are just three of them. One of my favorites is Here Comes Everybody by Clay Shurkey. There was actually a time when Clay was supposed to keynote a conference for which I was just attending and his plane got stuck in Kansas. And they asked me if I could just substitute with a talk of some kind. And I said I could, so long as we could name it, Here Comes Everybody except Clay Shurkey. So anyway, I cover the ways in which crowds can be enlisted to do interesting and helpful things. And I talk a little bit in the book that Martha mentioned, the future of the internet and how to stop it. This is the original cover of the book proposed by Yale University Press. I confess I was not that excited about it, particularly because my co-authors and I also had another book coming out the same month with a large hand saying stop. So we sent it back to them to do a new cover, and this is what they came up with. And I respected sort of the 8-bit font and the large traffic light, since after all the word stop was in the title, but didn't like that either. But instead of just saying no, I figured why not take a little bit of the own medicine that I'm preaching in the book. And I actually went to a website called worth1000.com as in a picture is worth 1000 words. This is a site where it turns out lots of people have Photoshop skills, and they go there because they have nothing else to do. So they participate in contests with one another to use their Photoshop skills. So I just chartered a contest called Please Create a Cover for my book. And amazingly, I got over two dozen entries in just a few days, and this turned out to be the winning entry from a guy named Ivo Van Der Int in Holland, whom I've never met and never will. And I did look at it and I thought it's a little bit depressing. It does have this kind of depressing quality. So I wrote them and said, can you just tweak it a little bit to give us some hope? I'll give you $50 over PayPal. Deal was accomplished, offer an acceptance, and that was the resulting cover of the book. So another very kind of positive use of the wisdom and activities of crowds. But then like a good academic, I started to get nervous just on general principles and tried to think through some of the worries that might come from exactly this kind of powerful instrumentality. And for that, I just tried to kind of organize in a simple pyramid what are an emerging range of platforms by which just sitting at your computer, you can start enlisting the brains and activities of distributed masses anywhere in the world. And at the top of the pyramid, I made it narrow because these are tasks that might be highly specialized or really require a lot of skill by the people you're hoping will do them. So that narrows the number who can participate and it might mean that you'll have to pay comparatively more money. And then as you go to the bottom of the pyramid, there are tasks you could imagine farming out this way that are so simple, nearly anybody with a pulse can do them. And the price through supply and demand gets comparatively low. So low in fact that it might drop to zero or even below zero. So what I wanna do is give you a quick tour of this pyramid from top to bottom and then share some concerns about how it might be working. So at the very top of the pyramid, we might put something like the X Prize Foundation. The X Prize started by a number of Silicon Valley successful entrepreneurs. They wanted to see good things happen in the world but to do so entrepreneurally. So they set for example a $10 million prize to the first group that could take a space vehicle, get it launched into suborbital space and bring it back safely in one go. And a number of groups competed. In fact, they found that over $100 million was invested by various groups trying to win the $10 million prize, which is great for society. It's good return on your investment, not as good for the group that came in second and didn't win anything. So that's sort of an example of a really tough problem. Another good X Prize is to design a simple car that can get over 100 miles per gallon. That's a progressive X Prize and a couple teams have stepped forward to win that. Maybe a little further down the pyramid, we might put something like Innocentive, a marketplace begun by Eli Lilly, the great pharma giant. They wanted to create a place where firms could set out difficult scientific problems. Not exactly Nobel Prize worthy, but still very hard. Here's how they portray it. Companies represented inexplicably by armless people send money over to scientists over here who then solve the problems and get the money. So it's this vernacular of seekers and solvers. So here's an example task. For $20,000 with a deadline of October 1st of just a year ago, please solve the problem of browning in juice. It turns out that maybe at a conference like this or a reunion, that when you go outside and there are some drinks waiting, there's those little tropicana bottles of juice and it can turn brown. Still be completely drinkable, but brown. So the state of the art in dealing with this right now is to cover the bottle from head to toe in a label so you don't notice that it's brown until you pour it out. Surely we can do better, says an anonymous company. And $20,000 goes to the first group that can make headway in this problem. And you can see over 270 people have started so-called project rooms online working with this anonymous company and they themselves are anonymous to solve the problem. And if they do by the deadline and they win, they get the $20,000 in all rights, title and interest, you can imagine how thick the contract is. In that solution, go back to the company, including the fact that the person even worked on the problem. This has been an amazingly successful marketplace where a lot of scientists with great technical skills and nothing better to do are making money this way. Maybe one step lower in the cloud, we would put LiveOps, your so-called contact center in the cloud. And the idea here from LiveOps, well here's how they explain it. In 2005 when Hurricane Katrina roared through New Orleans, the Red Cross advertised a toll-free number for people to call for information if they needed help. And of course it was swamped immediately. They got in touch with LiveOps, which within 45 minutes was able to bring on board thousands of extra people to answer the phones and to follow the script and to redirect people to where they needed the help. Now how did they do that? There weren't people waiting in a boiler room for a hurricane to hit. Instead, LiveOps has a little automated runway. If you want to work for LiveOps, you actually take a series of tests, kind of like the LSAT, to figure out just if you're a worthy or not. It tests your internet connection, how well your headset works. And about 3,000 people a week start down this runway, about 30 fall out the other side, ready to be employed by LiveOps. And when you're employed, what you can do is, this is how they describe it, calling all mompreneurs. So here you are in your house, there's your Diet Coke bottle, a bassinet is apparently over here, and you just plug in once you're a contractor and maybe you'll be taking orders at a local pizza joint. Not local to you, but local to someone else who's calling what they think is the pizza joint. You answer, you take the order, you type it in, you hit send, it goes back to wherever they are. The pizza gets prepared, everybody's happy. And if you do that well enough, you can sort of like in a video game level up. So level two might be taking orders at a fast food restaurant near the Squawk box. So you are the voice that comes out of the box when the car is idling outside. It might be traveling thousands of miles to you and you can calmly upsell the patron on an apple pie because you don't have to worry about the fry later overflowing. So you kind of do that, the order goes back to the restaurant, they drive just around the corner and pick up their meal and so on and so on. Very, very interesting model and one that is incredibly liberating to people that might have kids at home, might be wanting to work in very unusual hours. They don't even know what their hours are until they have them. You can do a few calls, take a break and then kick back in again. Now, maybe lower in the pyramid might be something, excuse me, like soma source. Micro work for the next billion. And the problem posed by soma source is if you have people in a refugee camp far away and all they have is a mobile telephone with the old variety, is there anything you could possibly display on that phone's screen and buttons they could press on the phone that would be of use to someone in the world who might be willing to pay a penny for whatever that person can do. To understand some example tasks that might actually fit that description, we turn closer to home to a very quiet corner of Amazon called Mechanical Turk. Mechanical Turk is designed to fill the niche of very small problems that you would think by now a computer could have solved but disappointingly, like jet packs, we don't have it yet. So that's why they call Mechanical Turk artificial, artificial intelligence. It's actually like a computer but there are really people inside who do the tasks. In fact, the name is coined after the famed Mechanical Turk of the late 19th century, pictured here playing a mean game of chess. You might ask, how did it play such a good game of chess? Well, it turns out there's a person inside the Mechanical Turk making it work, quite possibly the world's first digital sweatshop. Now, you may ask why a company would want to hold this out as its idealized model of the future of labor but it has been succeeding. Lots and lots of people line up to perform so-called hits, human intelligence tasks like this one, provide related keywords for these images. So for a penny, you are to provide a keyword when you look at this image. And it's funny, because as you look at it, you know you can't help yourself. You're like, box, box. Congratulations, you just turned to penny. And you go over here, you're like, trucks. Yep, here's another penny. Like I can get used to this. It's like a slot machine with a pretty continuous payoff. And it turns out lots and lots of people want to do it including some of the early demographic studies show quite a good number of graduate students use mechanical Turk. Either because they are somewhat poor at the time they're in graduate school or I think it might be one of the only forms of consistent positive feedback they get. Here's another mechanical Turk task for three cents. Tell us the difference between vanilla and French vanilla. Do it in your own words. It has to be between 50 and 70. Tell us the difference will pay you. Who wants to know? We don't know. Why do they want to know? Don't know. Just answer the question and move right along, nothing to see here. This is I think it fair to say profoundly weird. In the meta category of mechanical Turk tasks, here's a task where Turkers, that's what they're called, were asked to write on a piece of paper why they do mechanical Turk, take a picture of themselves and then send it in for 50 cents. So here are some of the answers. I Turk for Christmas. I Turk to battle insomnia, I believe him. I Turk for drug money, just kidding. I don't believe him. And it actually makes an amazing mosaic because you really see the deep inside what feels when you are the seeker, the tasker, like just putting it out to a computer at any time of day and answers just flood back, you realize like a honeycomb of a beehive, there are actually people inside doing the work. Very, very interesting phenomenon that now has been taken just in the past few weeks through colleagues over at the School of Engineering and Applied Sciences to its logical conclusion. They have designed something called Soylent, a word processor with a crowd inside. Now again, we have a huge spread of ears here. How many people know the meaning of the word Soylent? Anyone? A few. Soylent Green. So, well, you should have done a spoiler alert first. Of course what the gentleman is referring to is the movie Soylent Green with Charlton Heston pictured here looks very exciting. There's Charlton Heston running away and the movie basically has these little food pellets called Soylent Green that everybody really likes. And now here is the spoiler alert. If you're planning to watch this movie, cover your ears because at a key moment, the private eye, Charlton Heston makes a key discovery about Soylent Green. People, Soylent Green is made out of people. Again, you wonder about the metaphors people in this space use exactly, right? This is now Soylent, a word processor made out of people. So how does this work? It turns out you have Microsoft Word. There's an add-in you can get called Shorten and you highlight a paragraph and you click Shorten and it will actually shorten the paragraph in an intelligent way. And you can actually see it gets shorter as you go. And for you, it's just clicking a button in Word. But there's people out in the world making it happen through a somewhat careful process so you're not just trusting a single random person to make good edits. Now of course, because it's having to filter out through people who at that exact instant are taking up the job, it takes about two minutes to shorten a paragraph. Guess how long it takes to shorten 10 paragraphs? Two minutes is exactly right. I knew we had a good LSAT crowd in here. Two minutes because no matter how many paragraphs you highlight, it breaks them up and just sends them to independent groups. So it's always two minutes away from a good shortening no matter how long your document is. Now of course, if you happen to be general counsel of a company and you realize that people in the company are shortening their documents using this, you may be somewhat concerned about things. But put that aside for a moment and realize if you're worried about what if the paragraphs or something material say to the SEC that maybe shouldn't be going out to random people, they've got you on that because they're coming up with ways in which for instance if you were to take this handwritten form and need it to be converted into text in a reliable way, perfect task to send out to Mechanical Turk, well there are companies that will break the firm into individual chunks. So one person will just take the handwritten big spring and make that into big spring. Somebody entirely different will get the zip code, classic compartmentalization so that stuff is so broken up you have no idea what you're working on or for whom it's just translating, imagine translating one sentence at a time of a document from one language to another. Now the company that actually offers this service called MicroWork also looks forward to the day when instead of paying a penny to have this done why don't you pay nothing? Now how can we make it work that you pay nothing? And for that we turn to the work of a guy named Louise Von Ann, a Carnegie Mellon who in 2006 pioneered something called the ESP game. This is the same image labeling task that we saw from Mechanical Turk where we had the trucks and the boxes at a penny a pop. Here he just makes a game online for people to play where you are shown a picture and somewhere else whom you will never meet somebody else has shown the same picture and you try to guess what the other person is guessing on the picture and when you both guess on the same label you win points. Now the points accrue down here in this sideways wrong way thermometer. What are the points good for? Absolutely nothing. Kind of like frequent flier miles if you think about it. It's just blackout dates, forever. And you look at that and you say well who would possibly do this? And it turns out lots and lots of people do this. He was able to generate 4.1 million labels with 23,000 players some of whom play for over 20 hours a week. In fact his thesis advisor asked him to tweak it so that anybody playing coming from a domain that ended in .edu would be capped at 20 hours and told to get back to work on their thesis. Now I'm not sure I would conclude from that that the ESP game is fun but it is certainly compelling in a certain way and lots and lots of people do it. So he just runs the math and he says 5,000 people playing simultaneously could label all images on something like Google in 30 days. Incredibly powerful tool. Google took note, licensed the game. Luis now has a side business called Guap Games with a purpose and he's got lots of different games you can play. The thermometer is now three dimensional and here's one where you're tagging a tune and when you get something right you score again these points that are good for nothing. I scored 60 points here on this one thing. The most points that day, JC with 36,000 points. People really like points. We don't need to tell a room full of Harvard Laws graduates that, right? Sorry. So we've taken this to a conclusion that actually makes the initial pyramid I was talking about somewhat upended. Some researchers have come up with a way of deploying a game for so-called EDA, Electronic Design Automation. This is the kind of task by which we try to cram transistors on a chip ever closer together and it's really complicated to know how to do this and in fact we often have computers try to design their successors and they run into trouble because there's just too many permutations and it's really hard to figure this out. So these researchers have taken problems in that design automation and mapped it on to this very simple game where you click these rectangles between yellow and gray around the periphery and that makes these circles go between green and gray. And if you do it in just the right sequence they'll all turn green and then you win which will mean that you have solved a problem in electronic design automation that even the computers couldn't solve. They don't know ahead of time what the right combination is. You will discover that along with everybody else as you click and click and click intuiting what seems to be getting things lighting. So now we are at a law school after all so let's shift to a good old fashioned hypothetical. It's a Saturday morning. Your child or grandchild is happily occupied at the computer playing a game here at the PBS Kids site. These are the games you would normally see at PBS Kids so just one tweak for our hypothetical. Here's the game your kid is playing happily. Does anybody have a problem with that? So far no. Anybody not have a problem with that? Few people. Anybody just sort of bemused? Okay, this is clearly a room full of lawyers. Answer unclear but memo to follow. All right so hold that thought. I will come back to this question in a bookend and we'll see if we still have the same distribution of votes. In the meantime, let's look again at Innocentive. These are the people about browning and juice and here's another Innocentive task. They are seeking somebody pyrazolopyridinodiasonines. I don't know what they are. I have no idea. Just imagine the possible things that pyrazolopyridinodiasonines could be such that you might think this market has suddenly gotten dare I say well I'd get struck by lightning from Olin if I say this too efficient. Is it possible that creating arms length transactions in which people can generate mysterious chemicals that are otherwise hard to find on the market and send them to unknown the seekers might be to our detriment? It's kind of puzzling just how much we can now come up with tasks for which more and more of the people executing the tasks don't know for whom or why and we could think of some nefarious tasks where we wouldn't want that to happen. Now shifting away from just the point of view of the person doing the task, we can start to see some societal implications as well. So this is version 1.0 in 2006. Governor Perry in Texas, I think it was Perry set up webcams just along the border, put them on the internet and asked people to watch them all day long to make sure no one was getting in. A lot of people sat and watched the webcams looking for trouble. In fact this has been taken to its logical conclusion by a new UK company called Internet Eyes. As you might know London has the densest population of closed circuit television cameras per capita in the world, most of which are in private hands and which just record stuff so that if there's trouble you can review it later. Lenny comes in and looks at it and then you find out who did something wrong. Well this startup says, what if we could just have people watch the cameras live and report trouble as it happens? So here's a way to become a customer if you have a camera. Here's somebody relaxing on the couch staring at a camera who can earn reward money and in classic British understatement have a chance at reducing crime. The Thames Valley police when I was at Oxford they have emblazoned on the side of their car where an American police car would say something like to serve and protect or we'll kick your butt or something like that. Their car said reducing crime disorder and fear. I just wanted to walk up to an officer and be like sir I just want you to know I am extremely scared but not as scared as I would be without you. Thank you very much. Very strange. So anyway this is the updated version of the site they have new stock photography where she's now much more active looking at the machine and here viewers can register for free with no recurring fees. You don't have to pay to whitewash the fence. Isn't that terrific? In fact what happens is you stare at it and maybe a whole honeycomb of CCTV like bingo games at the church and you're waiting for something bad to happen so you can earn your bounty for reporting it and every so often like they do with TSA they'll stage a fake video of something going on to see if you're awake and if you catch it you can win points. So very very interesting model and we'll see how well that turns out. Now here's another 1.0 early version 2006 at the University of Colorado on April 20th a date that might have significance I believe for the members of the class of 2005 or perhaps those from the 60s. April 20th turns out to be a key date at which those kids gather on Ferrand field and smoke something they're not supposed to smoke and their idea is there are too many of us for the police to arrest. You can't arrest us all. The police said you're right but we can take pictures of you all and then put them on the web at $50 a pop for who can tell us who this person is. Now the reaction is somewhat interesting because it is the classic wanted poster. It's just somehow but you know the cat is only supposed to be this big relative to the mouse. It's fair to have a cat, it's fair to have a mouse but when it gets too easy for the cat somehow we object at least if it's over something like this and sure enough we've now seen version 2.0 from 2009. This is a real website translated from the Farsi that the government put up in Iran after the disputed elections for which there were protests and here there's a bunch of people they can't identify for which they have importuned the public to name who they are. Now of course there's a ceiling on how effective this can be because the people most likely to know who they are are probably the least likely to wanna turn them in because they know what's going to happen. So then I got to thinking how could we turn this hypothetically into a mechanical Turk task? We can't put it out and say we'll give you five cents if you know who this person is because in Topeka or Beijing or Bangalore they're not gonna know who it is but if you're the government of Iran you could just take some of your 57 million ID card photos, put them on one side, have the protest you're trying to identify on the left and just lather rinse repeat. Do a task where you say do you see the person here in any of the photos here? And I just did a quick kind of back of the envelope spreadsheet on it and found that basically for under $17,000 roughly you can identify any arbitrary person in a population of 57 million. Yeah, kind of weird. So here comes the bookended hypothetical which I suppose you have already anticipated. It's a Saturday morning. Here's what the kid is doing. Having a great time. How many people have a problem with that? More. How many people do not have a problem with this? How many people are bemused? All right, still a good number of bemused people but we pulled a number of people over into they have a problem with that side. And of course like a good law school hypothetical it's pretty much pulling at every possible string it can. It's kids, it's PBS, it's protestors. But that's an example I think of where you can see it happening, where you can see the market being created that really is an invisible hand that will push to make it so without any other intervention. Another set of worries. This is a real mechanical task where you are asked to review a website called natmedtalk.com on your blog and you have to have a very real blog and if you do, you review the site, you include a link, you can't disclose that this is a paid review and then you'll be paid $2. And of course it's the NatMed Talk people wanting to salt the web with good reviews of its site. That doesn't actually demand that you write a good review but an odd feature of Mechanical Turk is that payment is utterly at the discretion of the seeker. So if they don't like what you do they just don't pay you. And what are you gonna do? Litigate it over three cents? You could be earning three cents, 10 seconds later by not litigating and just doing another task. So I don't know, maybe somebody could say class action and in which case there's a very usable idea coming out of the lunch for somebody here. Here's another example which is a little clearer about wanting a good review. Write a five out of five review for a product on a website. Write as if you own the product and are using it. Tell a story of why you've bought it and how you are using it. And you will then earn, I think this was a 50 cent or $1 hit. Now, this was a real one. It turned out it was for Amazon.com for a product there which is kind of funny that here's one part of Amazon subverting another. And it leaked onto one of the technological bulletin boards onto Slash Dot. So you ended up with the Slash Dotters all rushing to the website to leave one star reviews. The Turkers all loyally leaving five star reviews. Nobody had bought the product at all. Now, I don't know how worried this makes you. The fact that any of these rating schemes, here's another called Dig where you rate stuff that you like can be subverted and there are now marketplaces in this. Here is a website called subvertandprofit.com. And you can go there and just buy any number of digs or thumbs up or Facebook likes that you want not approved by the respective sites. This is an intermediary who takes half the money and gives the other half to the public out there that's already in the business of digging things or liking things and just gives them 50 cents in order to do that. Now you might say again, hey, I don't trust what I see online anyway, so who cares? My thought and originally hypothetical was how long before this gets out of the virtual world and into the real world? So there's classic lobbying campaigns, this one for the European Parliament where we importune people to believe what we believe and then write their representative who will respond to enough people genuinely pressuring them to do something one thing or another. So my worry was, well what happens when you start using these systems to just pay people to lobby in the real world? It turns out this is no longer hypothetical. Some of you may have heard of gethealthreformright.org, one of the insurance funded organizations, I don't know if you'd call it grassroots or astroturf, but one of the organizations impaneled to fight healthcare reform and they just did the usual thing trying to get people to agree with them and then write their member of Congress, but then they took it a step further. They went into a game called Farmville which is wildly popular, has tens of millions of players around the world and they actually paid Farmville people in virtual crops in order to lobby their real world members of Congress. And that's the kind of thing for which the only thing that would make it a more perfect textbook hypothetical would be if this had been over the Farmville. So you'd have people getting paid virtual crops to argue against real crop subsidies. Somehow that would just be a perfect textbook example. This is the kind of thing that starts to make me extremely worried because the endpoint of it is, if you don't like something, you're upset with the Sheraton Commander for some reason. I don't know, maybe you're the Harvard Square Hotel. You go online within 20 minutes and a few clicks and a small check, you can have a flash mob of protesters waving signs that you have already specified, arguing against the Sheraton Commander and driving away business. And it's very difficult to know whether they are speaking from the heart or from the purse when they come and do that. Again, a great example of an extremely efficient market that can be leading to real trouble. So that's why I look at this original experiment from Casey Kinzer and I say, it's so amazing what you can get people to do, but I also see the other side to it and start to worry about what you can get people enlisted to do without them even realizing the mosaic that they are creating. Now, for every good that might mean a bad, for every bad a good. This is a website set up by a guy named Ben Rigby from a group called The Extraordinaries. The Extraordinaries is like mechanical Turk, but for nonprofits. If you're a nonprofit organization and you have small tasks you want help with, you can put it out for free and people help you because they want to help out a nonprofit and they are called The Extraordinaries. He actually was talking to me and I was telling him about the Iranian election hypothetical with mechanical Turk to identify protesters and when the Haiti earthquake hit, he turned it around and he actually created a site that people could go to to submit a photo of somebody missing from the earthquake. He then got a feed of images taken after the earthquake and asked anybody to sort through the images and see if they could find a match. He got tens of thousands of people to actually undergo this exercise at the site and identified over 20 leads from pictures of missing people that had been pictured as being found at some point after the earthquake. So that's kind of a hopeful sense of things and then I kind of end with this wonderfully ambiguous mechanical Turk task. You are to do something kind and then take a photo of it and it's pictured here by this oddly configured kind machine. People go in this funnel, gears, process them in hearts, come out the other side. Again, I just don't know what it is about the metaphors in this space that they have to be so lurid but in the meantime, the actual task is that you are supposed to do something nice in the world. Just go and be friendly, take a photo of it, tell a story and you will get 50 cents and people go out into the world and do this and you realize we have at last created a processor that converts money into love. Who knew you could do this? And it's like, all right, our work is done, let's just call it a century. And then on the other hand, you're like, the next time somebody does something nice for me, are they being nice or are they going around the corner to collect their points? It's kind of a weird thing that suddenly everything is able to be turned into points and it's that ambiguity that I think we ought to be working our way through and figuring out what kind of world is the world we want to see built with these new technologies and how we might be able to sways, regulate, otherwise see that world come about. Thank you very much. We have time for a couple of questions. So I think we have some time for questions and reactions before we break. There are microphones ready to roam the room. There's one here from our friend at George Washington University. The question was, how much is this points versus cash? And right now there are a ton of parallel experiments going on in both. And the discoveries, this is a lot of work on motivation and there are people here like Yochai Benkler working on cooperation and motivation and they've even come up with a vernacular of internal, intrinsic versus extrinsic motivation. What you do because it makes you feel good versus what you do because somebody's writing you a check to do it. And they have found that there are some amazingly powerful hybrids where people can be enlisted a little intrinsically but then stick around for intrinsic reasons. Or the other way around, they come because they believe in it but then they start getting a cash flow and then they depend on it. I think it's too early to tell which will be the right formula. But in Silicon Valley right now, a lot of the talk is about a phenomenon they call gamification where they find that turning almost anything into a game where you score points or lose points, even if the points are merely just glory and competitive turns out to alone be a huge motivator. Now whether people when they find that their points are helping somebody else out epiphenominally, whether or not they still want to see that take place is a good question. But so far they don't seem to mind. And if you think about some of the sites if you were picking a hotel, you might have used TripAdvisor or Yelp or something. That's an example of businesses whose worth is entirely premised on the goodwill and labor of unaffiliated people around the world who submit reviews. And they do it either because they want to do it or you can be designated Yelp Elite if you leave enough and you become a docent on Yelp. And you see yourself as sort of working for the man, but still it's fun. Other comments, questions, reactions? Over here? Yes, right over there. Like a lot of people, like a lot of people I've got a blog. It's about wine. It's read mostly by my mother. How much do you pay her? So increasingly the hits that I get every day, I check to see come from all over the United States, but they're people that stay 0.00 seconds, so they're not reading it. So what's with this? Well, if somebody's staying for 0.00 seconds, chances are good it's a so-called bot or robot. And it turns out that a lot of the entities crawling around the web are robots. Either for perfectly good purposes, Google sends its robot around continuously so that it can update its index so we know what's where. And it also turns out that a lot of the current internet economy is premised on banner advertising and other displays for which sometimes merely exposing an eyeball to it creates a money flow. Somebody looks at the ad, the person chartering the ad has to pay for that. And sometimes, this is the Google model, if they click through, if they actually generate a click on the ad because they're interested, that generates money flow. And that will sometimes mean people send out robots in order to click on stuff in order to get money flow going in their direction because Google in turn has affiliates. On your wine site, you could put up so-called ad words and if people end up clicking on the ads for the wine, Google shares the money with you. That gives you an interest if you weren't the upstanding person that we know you to be, you would click on the ads and Google would just start paying you for your own eyeballs. So that's a lot of what can account for this, but there's also, I almost think of it like in physics, they have dark energy and dark matter. And I was actually told by a cosmologist that they call it dark energy because they wanted funding from the Department of Energy. So otherwise they're not exactly sure what this stuff is. But there's sort of this dark energy out there on the web where whenever you try to do research and gather even basic demographics, like how many people are using the internet right now? Which would be a trivial question for a comp you serve of the earlier era to answer. How many subscribers do you have? Or for Verizon to tell you how many people they have on their mobile phone network or on their landline, it's so difficult to answer from the internet because we don't even know who's a robot and who's not. And that's why I can only give you a partial set of suppositions about those ever so brief visitors to the site. Maybe one more question, yes? Oh, and I see another back there too. So we'll do here and then we'll do the corner of the room. Thank you, Professor. I am sure that Dean Minow was very on the point when they say you're very entertaining and very amusing. Thank you. That's just a sign of intelligence. My question is not really on the talk itself but derived from it. What can we do as users to avoid getting unwanted mail and also to avoid from having our mail sometimes be classified as spam? Yeah. I'm sick and tired of getting Canadian pharmacy ads for Viagra spam and the Seattleis. Okay, I'm 60, I'm 70 and maybe I don't, I'm not confessing to anything. I may need it, I don't know. But I don't want the ads. I don't need the ads. But at the same time, I have had mail that I sent that's returned because they considered spam and it's called something, spam something. What can we do practically about it? Yes, yeah. So this question falls into the news you can use category and of course any answer has a shelf life of only about a week but here's a framework. And actually Dean Minow had mentioned the spam works article that I'd done with somebody named Laura Frieder at Purdue at the School of Management there where I was so frustrated with the amount of spam I was getting. I think there's an academic project in this, right? Make some lemonade out of these lemons. And we started to look at the stock spam, the spam that says buy now in such and such. And so we gathered as much of that as we could at various so-called honeypot places on the web where it just attracted the spam and then looked to see if we had acted then or being the spammer had the foresight of a day earlier been able to act on what the email was suggesting would we consistently be making money and we got a big pile of pink sheets data because these were largely pink sheet stocks being traded and we found out that the answer was yes you could consistently get a five to 7% return if you were aware of the fact of the spam going out just before it. And that just illustrates now across all spam not just stock spam, somewhere somebody is clicking on this stuff and the economics of it is so unusual that you can send out a billion messages and if two people click it's a good day. It's like you won as the spammer and the internet engineers had an unusual reaction to this. Their view was pretty much at all costs keep the network open, let the traffic flow and let the end points deal with anything unwanted which is to say it's up to you to filter it out or to have an internet service provider or email provider who will do the filtering for you. And there's good reasons to wanna see it happen but it's very counterintuitive especially now that we know through some testing demographics that easily over 95% of all email traffic is spam. If you just dip a ladle into the world's email 95% spam comes out and again internet engineers are like okay we'll just filter out that 95% and you've got 5% that's still the chicken McNugget and like that's what you like. That might make sense in network economics terms because even 10 seconds of high quality video is worth thousands and thousands of standard emails so from a network management standpoint it might be okay but what we have seen cropped up through this ethos are a range of so-called private sheriffs. Entities not chartered by governments that interpose themselves as sort of cartels made so by their success where you lean on them to do the filtering for you. So maybe you pipe your mail through Gmail because they do pretty good email forwarding or one of any other number of things and there was actually a guy who started this called Paul Vixie who just kept a list. It was Paul Vixie's like Santa Claus you've been bad list of entities that he thought had sent spam and if you got added to his list he would not listen to you he would filter out mail from you and he made his list available in real time to others who might want to also filter on behalf so Hotmail from MSN subscribed to Paul's list and that meant that if Paul didn't like you your mail was not getting to anybody at hotmail.com and if you complained to Paul about it he wasn't answering because he filtered out your mail and I actually had Paul as a guest in Langdell North and now the Vorenberg room in a class to talk about this new kind of lawmaking by private sheriffs in cyberspace and he was surprisingly diffident about it he was like look it's just somebody has to mop the floor I wish government would do this but since they haven't I'm stepping up to the plate which really surprised the libertarian professor who was another guest who was ready to say how great this was and this was the wave of the future it turned out that he had lots of competition one of them was called Orbs the open relay behavior modification system and Orbs would send out these test emails and if your email server were configured the wrong way you would end up on their bad list and none of your mail would go anywhere and the Berkman Center here at Harvard got listed by Orbs at one time we got an automated message that says you have an hour to fix your mail server or you may find that none of your mail goes anywhere and you know we did it and I asked Paul about this and he says yeah Orbs doesn't really have the human touch they don't get on my list until I say they get on my list you know they don't say Haynes till I say they say Haynes and he says a matter of fact I think those test emails that Orbs sends out I think they're spam I said what did you do about it he said I blacklisted them I said what did they do about it he said they blacklisted me surely there's a better way but in the meantime it's these private sheriffs that lead to the double-edged problem you see trying to shop for the right solution so you avoid the spam and at the same time as a sender of email having a very complicated matrix like Goldberg B. Kelly looks easy next to this to try to get your due process back if you are not communicating with a person that you think wants to hear from you a last question maybe in the back here Disney's Penguin Club and Barbie's Playhouse are nefarious and that they're teaching young children to spend huge amounts of time on the internet to acquire points unless of course mom is willing to up her credit card ante for them to buy virtual stuff for their avatar there's a purely hypothetical question I can tell so I think the practical news you can use answer is they are here to stay and if you should go so far as to deprive your child all of whose friends are using these sites of the chance to do it your child will end up sort of a virtual panhandler codging access through friends who have more lenient parents and will actually be indentured to other of the friends for access through their penguins so as a purely practical matter I think Club Penguin and its progeny are here to stay as a larger matter what we're inculcating the kids with and the ways in which through Club Penguin we're having these points and then there's marketing with it and maybe the thing worse than having the points be worthless is having them be worth something that might actually be worse because now you have all sorts of marketing tie-ins and now they can buy candy with their points and all sorts of stuff I do however think it's just a new way of interaction and it's just it's that much more incumbent in the parental relationship that will turn out to try to get your kids aware of the dangers and effect of gambling or of a whole range of what will turn out to be kind of obsessive compulsive behaviors for which these sites are tailored exquisitely often to their particular personalities to keep them coming back and pressing the lever to get the little pellet so whatever you think about Las Vegas this is sort of what's here to stay and I think again collectively interdisciplinarily and disciplinarily as a parent we're gonna have to figure out how to solve that thank you very much