A (0:00)

Okay, everyone, good evening and welcome to this 2011 summer school public lecture. For those of you who don't know me, I'm Andrew Murray and I'm from the Law department and I'm the director of the law program in the summer school. And I'm here because, at least partially, some of what Professor Meyer Schonberger is going to be talking about is legally related. So that's what I'm here for. And my main role tonight is to introduce Professor Meyers Schoenberger, which I'm going to do in just a minute. Before then, what I would say is that the success of the speakers we get to come to these summer school lectures reflects in essence what the LSE is and what it offers. Being here at the LSE is much more than just having access to the full time staff and the facilities of the lse. What we can also do is through our reputation and through our global standing, we can get some of the very best speakers from anywhere in the world to come along and to give of their time so that you can hear from them, whereas you might not otherwise have had the opportunity to do so. And this is how we can bring along speakers of the quality of tonight's speaker. So I think what I should do is introduce tonight's speaker and then I will pass over to him and he's going to talk for probably about 45 minutes and then we'll take some questions from you. Tonight's speaker is Professor Victor Meyers Schonberger, who's from the Oxford Internet Institute. It's really impossible to say how good Victor Maerschonberger is because he will deny it when he comes across here to speak. He has a number of degrees from great institutions, including best of all, a degree from the lse. He has written, I think. Is it eight books? Is it still eight?

B (1:55)

We're on. I forgot. Yeah.

A (1:56)

I think he's written eight books and a number of of subjects relating to privacy, the Internet society, the digital world, etc. His main one is called Delete the Virtue of Forgetting in the Information Age. Copies are available to buy outside. He asked me to say that I think the main thing that he would probably want me to tell you about him is that if you go to his website there is about a four or five minute video of him in a MiG 25 at 89,000ft, which he might say is the best thing he's ever done. But then that's before tonight's lecture. So I think with that ado, I'm going to pass over to Professor Meyer Schoenberger. And let him instruct you in forgetting in the Information age.

B (2:50)

Thank you very much for a wonderful introduction. Indeed, it's my only claim to fame that I flew a MiG 25. And before I start, I entertain you with a little anecdote that is not on my website. So I was in this MiG 25, which is a Cold War airplane that flies two times the speed of sound and was built to shoot down an American supersonic bomb, but it was never built. And it, the one that I flew in, had the pilot sitting behind me. And this was a special retrofit of the MiG25 done for the training of the Buran Russian space shuttle that never flew either. And so that's why it was a two seater. And my pilot said to me, if you. If you hear me say eject, eject. And I said, is there anything I need to do, like press a button or so if that happens? He said, no, you eject. And I said, that's wonderful. But what happens if something misha happens to you, like a heart attack or so? You look fit and everything, but, you know, 89,000ft up, two times the sound, the speed of sound. I'm a little worried. What do I have to do? If you fall sick, how do I eject? And he looked at me and he said, you cannot. And I said, but what happens then to me? If you have a heart attack and I cannot eject, can I land the airplane? You cannot. And then I said, what happens to me? And he looked at me and he said, you're dead. Well, I didn't die. He didn't have a heart attack. And so I'm here to talk about. Delete. That is the anecdote that is not on the web. All right, so for something different then. Stacy Snyder wanted to be a teacher. By the summer of 2006, she had completed her coursework and was looking forward to her teacher's certificate. Then, from one day to the next, her dream was over. She was summoned to dean of her university and told that she would not receive her certificate. She would not be a teacher. Although she had the credits, passed the exams, completed practical training, she would not be given her certificate, she was told, because her behavior was unbecoming of a teacher. Her behavior. A photo showing her with a cap and a cup captioned drunken pirate. Stacy Snyder had uploaded this photo on her MySpace webpage for her friends to see and perhaps to chuckle. But the university administration found the photo and found the photo to induce minors to consume alcohol and thus to be inappropriate for a teacher. When Stacy was confronted with the photo by the university administration, she considered taking the photo offline, but it was too late. Her photo had been indexed by search engines and archived by web crawlers. As much as Stacey wanted the photo to be forgotten, the Internet would not permit that. Remembering instead of forgetting Remembering forgetting in 2001, Andrew Feldmar, a Canadian psychotherapist living in Vancouver, wrote an academic article for a journal. In the article, he mentioned that in 1965 he had taken LSD. In the summer of 2006, like many times before, the old and graying Andrew Feldmar wanted to cross from Canada into the United States to pick up a friend from Seattle International Airport. The US Immigration officer Googled Feldmar and discovered the academic journal article from 2001. Because Andrew Feldma had failed to disclose to the immigration office, although he never denied it, that he had taken drugs 40 years earlier, he was interrogated for four hours, fingerprinted, and then barred from entering the United States. Forever remembering instead of forgetting. Of course, you may say now Stacey's and Andrews cases are tragic, but at least in part because of their own doing. Had they not put information online, Stacy would be a teacher now and Andrew could still travel into the United States. Everybody, to paraphrase Swiss author Friedrich Duhrmacht, has to decide for her or himself what to make available online. What once has been put on the web is no longer forgotten. Really? Do we really know? Every time information about us is being collected, stored and made accessible? For most of us, Google is the search engine of choice. Millions of people around the world send more than 2 billion search queries to Google every single day. 2 billion. Google is showing them the way. Google also shows us what is being searched, where, when, and by whom. This is a page of what is called Google Trends that shows you were and how and how often the term Iraq in this example, is being searched for in which region, in which city by which language. Google can do this because Google stores and monitors search queries. Now, this can be quite powerful. Remember about a year ago we had a flu epidemic, the H1N1 flu epidemic. According to a law in the United States, when a doctor has a patient with a flu with flu symptoms, the doctor has to notify the center for Disease Control in Atlanta, a government agency that monitors the outbreak of the flu, and then tries to measure the flow of the epidemic. The problem is that the notification from the doctor to the center of Disease Control takes some time. So the cdc, the center for Disease Control In Atlanta only receives information a couple of days late. And when they model it with their software, they see the flu epidemic two weeks earlier, but not as it is at this moment. Compare this to Google. Google just said, let's monitor who is looking for H1N1 information because there is a strong likelihood that these people feel sick and model therefore the spread of the flu based on search queries. Now, here is how well these two data sources overlap. The orange graph is the official United States flu data, and the blue one is just the Google data. The Google flu trends. In other words, because we search so much with Google, Google knows a lot about us, including when we get the flu. Google can do this even for events way back, as I showed you, with the term Iraq, because Google does not forget. Since Google's humble beginnings more than a decade ago, Google has stored every single search query it ever received. 2 billion a day. Right now, every single search query it ever received and every search result you ever clicked on. Remembering, Forgetting for millennia, for us humans, forgetting has been easy. It's built into us biologically. We forgot most of what we experience every day. Our feelings, our thoughts. Remembering is hard. Since the beginning of time, therefore, we humans have tried to overcome biological forgetting and to hold on to memories that are precious. For thousands of years, we have tried, like this Navajo Indian, to pass on our memories to our children and the hope that they too may thus be able to remember. This is how the great epics in the world came into being thousands of years ago. But human memory is not fixed. It changes as we reconstruct our past. Depending on it, therefore, is not sufficient, especially when we want to capture something precisely or for a long period of time. Painting is one way of encapsulating visual impressions to create an external, a more precise and more lasting memory, like this beautiful cave drawing from the caves of Altamira. Script, originally developed, will you believe it, by accountants. Searching for a precise method of remembering has for millennia remained humanity's preferred external memory. Language, painting, script provided us with the capacity to remember across generations and through time. But these tools have not altered the fundamental fact that for us humans, forgetting is easy and remembering as hard time consuming, costly. The book did not change this. Neither did the phonograph or film. Remembering remained expensive for most human beings and was thus chosen carefully. In the 1930s, my father received a Kodak brownie box camera from his father. Photographs. Taking photographs, developing photographs was very expensive in that time. So my grandfather instructed my father, be careful. Make sure that a photo is really exceptional when you Take it, because every photo you take costs money. In the 1970s, when I was a small boy, I got a Kodak camera from my father. I don't get paid by Kodak for doing that. I got a camera from my father and he said to me, be careful. Every photo you take costs money. Reserve it for the real exceptional moments in life. In other words, all through human history, forgetting was the default. Remembering was reserved for the exceptions that enabled us humans to deal with time. Through our physiological capacity to forget, we rid ourselves of excess memory. What has been long past fades in our mind. Thus we pay tribute to time. We depreciate what is no longer relevant to our present. But because forgetting is biological, we humans never had to develop a cognitive method, a cognitive capacity to deliberately forget, to depreciate memories and to make them fade. Today this is different. Google remembers, Yahoo Remembers, Amazon remembers, the Internet Archive remembers. Flight reservation systems. Remember Flight reservation systems Remember? Yeah, flight reservation systems remember. If you look at a flight but never book it, they still remember for six months that you looked at that flight. Just in case law enforcement ever needs that, and not always do we realize that we have contributed to digital memory like with the flight reservation systems, and made them accessible to others? For example, millions of Facebook users, I'm sure including you around the world, change their profile when they begin or end a relationship. This not that you are in a relationship or not, but that your relationship status changes is accessible as part of Facebook's open graph and can be captured third parties in the aggregate. This may reveal amusing trends like this chart depicting when relationships end most frequently over the course of the year. I that is April Fool's Day's joke's gone bad. But such wonderfully funny data in the aggregate can also be analyzed on an individual level, indicating not just who you are in relationship with, but how long your relationships last over time, how many you have had and when. Or take another example. This is a map of London. This map of London is derived from photos that have been uploaded to Flickr. I don't have to introduce Flickr, right? 5 billion photographs now the iPhone is the most fragrance frequently used upload machine or photographic machine that is connected to Flickr. What this map shows is where people in London have taken photos and uploaded to Flickr. The map is possible because as you take photos with your iPhone or at a smartphone you can and upload to Flickr you not only transfer the photo up, you also transfer metadata, so called exif data up to Flickr for other people to look at and this EXIF data includes, in the case of the iPhone and many other smart devices, longitude and latitude. And the hotter these points get yellow or red, the more people have taken photos there. Now that's wonderful because with this map you know exactly where there is a good photo opportunity, so that your photo shots that you bring home from London are really great. On the other hand, so this is on the aggregate level, on the individual level. On the other hand, this might also give others a chance to follow and track where you have been when from biological forgetting, therefore, we have moved to a state of comprehensive remembering. How did this happen? You know this as much as I do, very briefly, four elements. The first one is digitization. That is that everything around us can be translated, stored and manipulated and then transmitted in a digital code of zeros and ones. Second, of course, are advances in storage technology. In 1965, a young engineer by the name of Gordon Moore surmised that the density of integrated circuits chips might approximate a doubling every two years. Importantly, if you look at this storage capacity, storage density tracks processing capacity over time. Very similar slopes. If you think that this is very impressive, that the slope is impressive, do have a look at the Y axis. This is a logarithmic axis. So if you would graph it on a linear graph with a linear axe, it would go of course up here. Now, Gordon Moore, who first surmised this in 1965 looking just at chips, not at digital storage devices, was very famous for his suggestion it might double every two years. It became Moore's law and he himself started a small hardware company that is no longer a small hardware company, but a chip company called Intel. But storage alone is not sufficient. The East German secret police, the Stasi, had hundreds of millions of facts in its files and almost a million of its citizens. Yet with its elaborate system of pseudonyms and codes and mostly paper based files, it had difficulties retrieving the information it had in time. This too is different today, as full text indexing, prohibitively expensive only a few decades ago, is so affordable today that it is not only what drives user expectations on the Internet. Think about Google Live, but it's also built into our major file systems. Add to this fourth, the ability to access information through a global digital infrastructure. A few minutes are sufficient to disseminate a document worldwide, even accidentally, and to have it distributed to millions of people, as this page from the manual of operating US Presidential airplane Air Force One, which was made available online accidentally for a very short period of time by the US Air Force. Once the mistake was realized, of course, it was too late. And by the way, these are the instructions. How to get into Air Force One, just in case you ever need it. Taken together today, this has led to remembering becoming the default and forgetting the exception. To an extent, this ought to be a reason for celebration. Yes, our vast and accessible digital memories offer numerous benefits. Increased accuracy, improved efficiency, all the way to the promise to help us transcend human mortality. At the same token, undoing forgetting has consequences, I believe, far beyond the narrow confines of information efficiencies. Two terms characterize what I believe is truly at stake. Power and time. Power Power is relative and relational, as information privacy scholars have long argued, by power over information may translate into power over the individual the information pertains to. But such informational power reaches far beyond the confines of dyadic relationships and information privacy. Consider this. For centuries, the Catholic Church rested its power in no small part on its domination of the institutions of remembering, from scribes and books all the way to libraries. A societal consequence of such power imbalances has often been for those that suffer from them to choose silence. And this is precisely what power holders intend. It also has the potential, of course, to influence how we interact and how we transact. But there is more to it. Take the idea that Jeremy Bentham had. Jeremy Bentham's panopticum is the concept of a prison. A prison in which the prison guards can watch the prisoners without the prisoners knowing whether and when they're actually being watched. The aim of the panopticon is behavioral compliance of the prisoners through the permanent threat of invisible surveillance. Oscar Gandhi and others have suggested the Internet may help create a global panopticon in which everybody has to assume that she is being watched all the time. Such panopticum may lead people to self censor, fearing that their utterances could be misconstrued by any of the hundreds of millions of individuals in tens of thousands of jurisdictions connected with the Net. But today, I believe we face more than just a global panopticon because of comprehensive digital memory. We have to assume that what we say or do today is not only witnessed in the now, but will remain accessible for years, perhaps decades. Remember Andrew Feldmar into our future. This creates a temporal panopticum in which we may self censor, not because we are afraid of how others might interpret our words and deeds today, but because how people and institutions and the future might view them. My second concern is time. Or more precisely, how we humans have dealt with time. As I mentioned earlier, forgetting is biological, so we did not have to develop conscious mechanisms to put different and perhaps contradictory pieces of memory in a temporal perspective. Let me try and make this more concrete for you. Consider I'm a lawyer at heart. Consider the following hypothetical story Jane and John Jane and John are old friends. Although they live in different cities now, they try to catch up at least once a year. One day Jane receives an email from John telling her that he is coming to town and looks forward to having coffee with her and to catch up. Jane is excited. She hasn't seen her old friend almost in a year and she is ready to write back right away suggesting a nice place to meet to remind her where they met last time. She's querying her mailbox folder up pop dozens and dozens and dozens of email messages that she has received from John over the years. She's quickly browsing through them, trying to find the right one, but then her eye catches a 10 year old email with a strange subject line. She starts to scan its text and then begins to read. Surprised, perhaps even shocked, Jane reads about how John deceived her and she's revisiting angry exchanges back and forth between them. Slowly, the events in her feelings triggered by this concrete external stimulus, the male message come back to her mind, her sense of betrayal and deception. She reads on about how over the following months and years, John and herself must have reconciled, although exactly how and when and why the emails do not tell. But at the forefront of her mind is now how John, her good friend John, deceived her and suddenly she's not so sure anymore she wants to meet with him when he comes to town. As much as her analytic mind tries to disregard the revived memory, the angry words she read triggered the recall. They are the external memory that help us remember things we thought we forgot, but they also, as in this example, may cloud our ability to evaluate and to decide. Put in more abstract terms, as cognitive psychologists remind us, for us humans it is difficult to realize time as a dimension of change that may trigger incorrect decision making. As we have just seen with Jane in the Analog Times, the danger was there, but it was limited. Our biological forgetting obscured our cognitive difficulties with time. But what when we are not permitted to forget anymore, we know a little bit about the consequences. Through studies of a handful of people who have biological difficulties to forget. This is aj, a woman who has difficulties forgetting. Ask her about a particular day and she's able to tell you when she woke up, what was on television, when what was in the newspaper, who called, how the weather was for every day the last 30 years. AJ cannot forget. But for her, it is not a blessing, it is a curse. She is haunted, she says, by her past. So much, in fact, that it limits her ability to decide in the present. She says that every time she has to decide in the present, she remembers all her failed decisions of her past. And that incapacitates her capacity to act. As Argentinian writer Jorge Luis Borges said, perfect memory pushes humans to get lost in detail, with no ability to generalize and to abstract and to evolve. They lose Borges rights. What makes us truly human, tethered to an ever more detailed past, rather than living and acting in the present. This is the fate we may face with comprehensive digital remembering. Through perfect digital memory, we also deny each other a capacity to change over time, to evolve and to grow without forgetting. It is hard, it turns out, for us humans to forgive. And so, with comprehensive digital remembering, we may turn into an unforgiving society. But there is one more wrinkle to the story. What if? What if, frustrated with the shortcomings of our own human memory, we begin to disregard our own recollection of our past and instead depend on digital memory? Does that give those that control digital memory, the Googles, the Flickrs, the Facebook of the world, the power to change history? These are some of the threats of shifting the default, I believe, from forgetting to remembering. So 70% done. What to do? For an academic, that's the hard part. So some responses already exist. The first and most obvious one is to enact information privacy rights. The idea of information privacy rights is intriguingly simple. By giving each and every individual a right to informational privacy, we empower the people, the individuals, to fight for their rights. Enforcement is both decentralized and delegated. Sounds great, but it comes with a number of inherent weaknesses. Most importantly that those that we aim to empower, the citizens, do not care. In Europe's strong information privacy rights have been enacted decades ago, but by and large, people have not used them. So maybe something else. Information ecology is a second approach. That is the conscious restriction, regulatory restriction of what personal information can be stored and for how long. Such information ecology norms necessitate government action and compliance and enforcement is costly. But they have two advantages over individual privacy rights. They do not require individuals to go to court for enforcement, and they protect against an uncertain future. Protect against an uncertain future. Consider the case of a Dutch citizen register put in place in the Netherlands in the 1930s for perfectly good reasons. To ensure the administration of Social Security, the register included information on religion and ethnicity. Once the Nazis had invaded the Netherlands. They specifically targeted the register, repurposing the information in it to identify Dutch Jews and send them to concentration camps. As a result, proportionally speaking, more Dutch Jews were murdered by the Nazis than those from France, Poland or Germany. Even Jewish refugees in the Netherlands fared better because they were not included in the Dutch citizen register. It's a horrific lesson. It's a horrific lesson because as we cannot foresee the future, and thus how personal information about us will be used, it may be better to store less than more. This is the essence of information ecology norms unfortunately, since September 11th, we have seen a major backlash here, together with a wave of information retention laws as part of a rhetoric of fear and security, thus limiting the political chances for a much needed expansion of information ecology to address digital remembering. So perhaps, dare I say, we need to think beyond laws. Some have argued for digital abstinence for staying away from the technical tools that enable digital remembering. Not sharing everything on Facebook, President Obama reminded us, may certainly reduce the threat of digital remembering. But is it realistic with over 700 million registered users worldwide? And would we want to deprive ourselves of the value of information sharing and peer production that the Tools of Web 2.0, including Facebook, provide us with? Another option? Ostensibly the exact opposite of digital abstinence is the idea of full contextualization or hold your breath to store digitally as much information as possible. Now that might sound crazy, but here's the argument. Perhaps the problem with digital memory is is that it does not capture enough of the context of an event to let us relive it later on accurately enough. So if we only could store everything, including the context of an event, we could avoid the negative side effects of digital memory. In essence, therefore, full contextualization would help us regain our ability to think in time. At the same time, it would also equalize any information imbalances at the core. This is Brin's proposal of a transparent society. But will full contextualization ever be technically feasible? How would we bring it about? And even if it were feasible, do we really have the time to relive all of our own past again and again and again and again, only to grasp what experiences in our past are no longer relevant to our present? That sounds like a waste of time. So perhaps a further alternative is to hope for cognitive adjustment in our society. That is to hope that over time we learn to devalue older information and to live in a world with an omnipresent past. Not society has to change or its laws, but our individual Process of information evaluation and decision making. That sounds right. Going into this research, I thought this is going to be the solution because it would certainly solve our problem. Unfortunately, cognitive psychologists are extremely skeptical of our human ability to force a change in how we evaluate and process distant memories that we suddenly recall through the external stimulus of digital storage. Remember the story of Jane? They suggest that it may take us humans a very long time to rewind our brains and to change the way we assess information, to modify essentially what we have been doing for ages. And what would be the appropriate mechanism to bring about such change. A different idea is not to change humans because they're so hard to change, but to change technology. Some have proposed to use technology to change behavior. We could, they argue, create a quasi property right in personal information, something like copyright, and build it into our technology, our phones, our PCs, our hard disks, and so forth, so that the technology would be built in to ensure that only those can process my personal information who I have permitted to do so. In short, the suggestion is to create a global digital rights management system to protect privacy. But wait a minute. Do we really need to create a global technological infrastructure that needs to watch every hour moves to ensure that nobody abuses somebody else's personal information? Would we not thereby create a perfect surveillance system in order to ensure privacy? I presented six possible approaches to deal with the challenges posed by digital remembering, privacy rights and information ecology. In the middle, employ legal norms to address the challenges of digital remembering. Digital abstinence and cognitive adjustment on the top. Hope that this could be achieved on an individual level, while privacy, DRM and full contextualization mainly rest on achieving technical breakthroughs. The three on the left target the power aspect of digital remembering. The three of them on the right aim at addressing the time challenge. None of these six offers us a silver bullet, although all of them help in their very unique way. Hence, we may need a mix. We need to combine them and perhaps even add something else. 95% done something else. I advocate for a revival of forgetting, that is to establish mechanisms that ease forgetting in the digital age and that make remembering just a tiny bit more strenuous. Not by much. I do not want to overly burden remembering, but just enough to shift the incentives of forgetting and remembering back to what we humans are used to. One version of this could be called expiration dates for information. That would imply that whenever we want to store information, for example on Facebook or in the cloud, we are prompted to enter not just say, the name of the file and the location of storage, but also a date until which we want the information to be stored. Once the data has been reached, the information is deleted from the system. Of course, we can choose an expiry date at will and we can change them anytime. While I want you to consider this, I do not want to impose more of this, my solution onto you at this point. You have been patient enough to listen to me and the core message that I want to convey to you is the challenge we face and the shadow of digital remembering. If you're interested, I must say, in finding out more about digital exploration dates in particular, but also remembering and forgetting in general, please take a look at my book Delete. But let me add as a bonus, one more example of how we could have a form of digital forgetting relatively easily. Remember digital photos and analog photos? Remember analog photos that you put in a box? Well, you can't remember. I can remember putting analog photos in a box and putting them in the attic or in the basement. I still have them. And if I want to spend the time to go up in the attic, I can still find them. Although some of them might have faded. But retrieving them takes just a little bit of effort, thereby forgetting impractical senses, the default and remembering the exception. That little bit of extra effort is what may help us stay focused on the present. Forgetting, remembering. Since the beginning of time, forgetting has been easy for us and remembering has been hard. In the digital age, that relationship has become reversed. Today we digital remembering is the default and it is forgetting that is often forgotten. I urge you, I urge you to give back to forgetting the role it deserves. Let us remember to forget. Thank you.