Technology & Science

Society needs to teach digital systems how to forget, author says

Interview with Viktor Mayer-Schonberger, author of 'Delete: The Virtue of Forgetting in the Digital Age,' on the impact digital technology and cheap storage are having on society.

Viktor Mayer-Schonberger, author of 'Delete: The Virtue of Forgetting in the Digital Age,' on the impact digital technology and cheap storage are having on society.

Viktor Mayer-Schonberger.
In his new book, Delete: The Virtue of Forgetting in the Digital Age, Viktor Mayer-Schönberger writes that forgetting is a natural human process, and that digital technology and cheap storage are creating all sorts of unexpected problems.

Mayer-Schönberger, the director of the Information and Innovation Policy Research Centre at the National University of Singapore's Lee Kuan School of Public Policy, argues that the result is an assault on privacy, an inability for people to make decisions, and a society where today's perception of reality is being coloured heavily by a digital past that is kept in storage and always at our fingertips.

Nora Young, host of CBC radio's Spark, interviewed Mayer-Schonberger to find out more about the digital tracks we leave in the course of our everyday lives, and what the digitized records of our activities may mean for our future if we don't learn to put them in the proper context. Here are excerpts from their discussion.

Nora Young:  At the beginning of your book, you write about the case of a Canadian, a man named Andrew Feldmar. Can you tell me that story?

Viktor Mayer-Schoenberger:  Andrew Feldmar is a psychotherapist out in Vancouver. His story was reported in the New York Times. He tried to go to Seattle-Tacoma International Airport to pick up a friend one day, and on the Canadian-U.S. border was Googled by a border guard. The border guard uncovered an article that Feldmar wrote in the late 1990s about his experiences using LSD in the 1960s.

Based on this piece of evidence, he confronted Feldmar with it and asked him whether that was accurate. Feldmar said yes, and then Feldmar, because he didn't disclose his illegal use of drug in the 1960s, was barred from entering the United States — not just for that one time, but forever.

Nora:   This is in a way a cautionary tale for you in this book.

Viktor:   What is particularly troubling about Andrew's example is the fact that the internet, that the system [does not] forget. Our digital traces are now remembered forever because the digital tools that surround us are capable of storing and retrieving information at very low cost. Because everybody can do that, now we, now the system, now our digital tools, remember and create a digital memory that is more comprehensive and perfect than any human memory could ever be.


Listen to the full version of the interview on the Spark website, or download the audio file.

Nora:   Yet that's at the root of so much of what we desire as human beings, to have photographic memory. What do you think the impulse is behind that desire for a perfect memory?

Viktor:   Well, the human impulse to remember is the fact that biologically we forget. Because we biologically forget most of what we experience, we try to hold on to, cling on to the experiences, the memories, the opinions, the values, the facts that we treasure and therefore want to remember. But the important element of that is that for all of human history, remembering has been hard and costly, and thus required active work and energy to be put into it, while forgetting was easy. Today with the digital tools, it's the other way around. Digital remembering is the default these days, and forgetting has become much, much harder.

Nora:   Can we flesh that out a little bit? What are some of the technological reasons why this incredible amount of memory is now at our disposal?

Viktor:   There are a number of reasons. One is of course the digital, binary code itself, the way by which our information processing systems work. Because all computers are general processing units, they can process any kind of information … whether it's sound, whether it's video, whether it's images, whether it's text, at least in principle. Through the internet, we can connect all this.

Then we have storage devices, the capacity of which have doubled pretty much every 14 to 16 months for the last 30 to 40 years. At the same time, cost for storage has come down. The net result is that … today when we shoot digital photos, we don't worry about shooting 10 pictures or 20 pictures. We go for 40 or 100. If we shoot digital video, very much unlike Super 8 reels, we don't stop after 15 seconds. We continue taping. Then we upload it onto our computer hard drives, and we are not particularly selective because the software that we use says, "Upload all of the pictures that you have on the camera?" and we say, "Sure. Go ahead, do it, " because 100 gigabytes, one terabyte, two terabytes — who cares? I have enough space on my computer to store everything. So we become the centerpiece of a network of digital tools in which the storing and the remembering has become the default and the forgetting has become the costly exception.

Nora:   You actually lay out how extraordinary this cheapness is, that it's actually not worth our time to even go through the photographs to decide which ones we want rather than just keeping them all, because to spend the time to decide is in a sense more expensive than just keeping all of it.

Viktor:   Absolutely. If we consider high-resolution digital photographs today, just the two, three, four seconds it takes to consider whether to keep that photograph or not on our hard drive is more costly than the hard disk space this photograph takes up on our hard drive. That's how costly it has become to forget, and that's how cheap it has gotten to remember.

Privacy issues

Nora:  So why is this a problem?

Viktor:   It's a problem for two reasons. One is connected to informational privacy, concerns that have been raised about the digital lifestyle for quite some time. It has to do with the fact that others might accumulate and collect personal information about us, and continue to store it and search it for a very long period of time, thereby knowing more about us than we can remember.

Just consider Google. For a very long period of time, Google stored each and every search query it was asked and associated a user to it. Only very recently, Google decided that it would forget search queries after nine months. That means that Google knows for nine months each and every search you asked it. That's more than we will remember about ourselves, about what we searched, and when, and under what circumstances, and which search results we clicked on. That means that Google has more informational power over us, and that's a concern.

A related concern is the fact that we must understand and realize whatever we say today will or might be held against us in the future. Others have described the Panopticon, which is basically a prison where all of the prisoners are potentially constantly watched, but don't know whether they're watched or not, and therefore they behave because they could be watched. The interesting thing is that perfect digital memory is not just a Panopticon. It's not just putting us in a glass cage and exposing us constantly to the rest of the world. It's creating a temporal Panopticon in which what we have said a month ago, or a year ago, or a decade ago will be held against us.

That's Andrew Feldmar's story. What he did 40 years ago and what he published maybe 10 years ago will be, and is, held against him today, irrespective of the fact that what he did 40 years ago has nothing to do and no relevance for the person he is today.

That's the second problem that I see: perfect, comprehensive digital memory denies human beings the ability to grow, to change, and to evolve over time. That is deeply worrying.

Jorge Luis Borges once wrote a wonderful short story called Funes, the Memorious, in which he describes a young man who, after an accident, cannot forget any more. He remembers everything. Borges says this is the person that can't see the forest for the trees, that has unlearned how to abstract, and how to generalize, and, therefore, how to be human. Now today there are few human beings who, for biological reasons, cannot forget. It sounds like a blessing, they certainly do remember where they parked their car in a shopping mall. [But] it turns out that they have tremendous difficulties in acting in time, in deciding in time, because they remember all their bad, failed decisions in the past, and therefore hesitate to make a decision in the present. Because they're forever tethered to the past, they can't act, they can't stay put in the present, and they can't imagine the future. I fear that with digital comprehensive memory, we might resemble these human beings, and we might lose our ability to act in time.

Human vs. machine memory

Nora:   Is that part of the reason why, right in the title of your book, you call forgetting a virtue, that idea that it's part of what allows us to grow and be creatures who exist in time?

Viktor:   Yes, absolutely. All the cognitive theorists, all the cognitive scientists who have studied brain, and memory and remembering, tell us that it is incredibly important for us humans to forget, because forgetting is a way by which our mind orders and prioritizes our experiences, makes sense of it, creates an environment, a context in our memory that makes us act. I am worried that digital comprehensive memory might undo this.

Nora:   What is the difference between human memory and what we call computer memory or storage?

Viktor:   Computer storage is the capturing of a flow of bits and the retrieval of a flow of bits. It is de-contextualized. It is taken out of its original context. We can only guess in what context, in what situation, it was said, or done, or captured.

Human memory, in contrast, is capable of putting things in perspective, of linking events with other events and experiences, of adding context and feeling to it, of reconstructing memory to fit better our own later perception and understanding of what happened. So it's eminently more versatile and eminently more useful, to be honest, than information storage in automatic information processing machines.

Nora:   This is sort of a personal bailiwick of mine, Viktor, but do you think we're starting to actually lose that distinction between remembering and recording?

Viktor:   Yes. This is a wonderful linkage that you have created. I haven't thought about it. I wish I had, because it is extremely powerful. What we do today, what we want our digital tools to do is not to remember, but to store. What we want us to do is not to remember, but to rely on and base our decisions on information storage. I think basing our decisions, basing our thinking, on human memory is eminently more useful than relying on information storage and information retrieval, which only, by the way, could record a certain slice of the context and the environment in which particular information and particular information exchange took place. We can only capture the words that have been said, but not the thoughts that were thought. That is something where only human memory can help us.

Nora:   Do you think, though, that this could be a cultural or a generational clash, that we are seeing a world that forgets and a world that remembers butting up against each other, and that maybe things will be different for the generation that's now growing up with this persistent amount of storage?

Viktor:   It's possible, but I am skeptical because if you ask young people who share a lot of information on social networking sites, and YouTube, Flickr, and so forth, they still are concerned about their informational privacy. They still are concerned about others gaining informational power over them, about being forced into a glass Panopticon.

The problem is, in a lot of circumstances, young and older people don't realize when they share information on the internet that this information not only is shared with potentially everybody, but that this will also remain accessible potentially for a very long period of time. Once we begin to become aware of these implications, once we begin to acknowledge and understand that digital memory is comprehensive and enduring, we may become extremely more cautious in what we do online. I don't like that idea. I do like us to share information and to use these Web 2.0 tools that we have that are so powerful in sharing knowledge, and experiences, and understanding with each other. I would hate for these tools to fall in disuse because we fear information sharing so much.

I'd also hate it if we become a nation or a world of self-censoring, where we don't want to share any thoughts with each other anymore for the fear that they might haunt us in years to come. What I want is a world that is teeming with information-sharing and information-exchange, of experiences being shared among people. But also a world in which we are aware that information is not endless, but has a life span, just like the yogurt in our refrigerator might expire over time.

Nora:   When we talk about social media, and the fact that we are the creators of a lot of the information online ourselves, we often talk about the Feldmar case, where somebody inadvertently reveals things. But as you say, the other side of it is this Panopticon idea, where you hear people talk about themselves as a brand, and you are presenting yourself as almost a marketed figure, because you know that there is this persistence of the image that you put out there.

Viktor:   Right. I'm encouraged by the fact that there are more and more Web 2.0 companies out there who offer information-sharing tools that have expiry dates attached to them. This for me is a sign that there is a market, and that people do want to share information, but also share information with a certain rule attached. And the rule says, "You can have this, but only for a limited period of time." Then it expires and vanishes, but until then, we can share something. And I like that idea, because it comes closer to what human memory is all about.

Seeking solutions

Nora:   Can we talk about that a little bit more? About what the solutions you see for this problem, this persistence of information?

Viktor:   Sure. The bottom line here is I think that people become aware of the fact that information is being stored, and as information is being stored, they become aware of the need to selectively forget information that has lost its value. We must understand that information is also linked to a certain point in time — most information, at least, is. And whether or not somebody got a ticket for drunk driving 25 years ago, when he or she was 18 years old, should have very little impact on how we judge that person today. That means information over time should lose its value, should lose its importance, and we need to build that into our digital tools and digital devices that surround us. We have built it into our own brain and our own cognition; in fact, it's built into it biologically through our capability of forgetting. Now we need to do the same thing with the digital tools that surround us, and we haven't done that yet.

Expiry dates, the proposal that I put forward, are one way of doing this, but there are many others. We can create rusting functions, so that information that is old takes longer to be retrieved, or can only be retrieved not completely, but haphazardly. So we can replicate some of the elements of human forgetting into our digital tools. Expiry dates are just one version of it. But expiry dates is an idea that captures the essence of making us aware and keeping us aware of the need to attach a time, a point in time to information. And to understand that information is not enduring forever, it is not eternal, but loses its value over time, and therefore should and must be forgotten.

Nora:  So how would something like expiry dates actually work?

Viktor:   Take search engines, for example. If you make a search request — that is, place a search query — that is automatically stored in the search engine's computer right now for many, many months to come. How about if you just have a button to press, and the search engine would delete all of the search queries that you have amassed, or delete the last 10 search queries, or delete a search query that you just entered maybe in a month's time? That would actually be advantageous, because if a friend came to see you, and you searched together on something that has no interest to you in the future, that particular incident or particular experience could be deleted out of your search history, and search results in the future would be more accurate with that.

Or take digital photography. You would take a picture, and instead of uploading the picture, you would ask for an expiry date for each of the pictures. You would have easy-to-use presets and defaults, so that it wouldn't take much of your time to enter the expiry date — perhaps a second or two for each photo.

And you could even, if you are particularly paranoid, have a software tool that alerts you if and when a particular file or image comes close to its expiry date, and you could change it if you want to.

All that this is meant to do is to raise the awareness and keep everybody aware that information is not endless, but information is tied to a moment in time. And sooner or later, it's a good idea for most information to be forgotten. Now, the cornerstone of expiry dates, and one that I favour a lot, is the fact that individually, as well as a society, we can choose the expiry date. We choose how long we want to remember, or when we want to forget. It's not our digital tools that by default make that decision for us, or some government agency, or some large corporation.

Nora:   Viktor, another character in your book is a researcher at Microsoft named Gordon Bell, whose goal is to digitally record his entire life. What do you think of his project?

Viktor:   I think Gordon Bell's project is wonderful, because he is creating his own digital treasure, his own digital memory of his life. It's a wonderful experiment. But there are two elements to keep in mind. One, Gordon Bell is adamant that he doesn't want anybody to have access to that information. It is only for himself. And therefore, he is very much aware of the privacy and power implications of this. Secondly, Gordon Bell doesn't tell us what he is using this archive for. Is he actually, every time he is facing a decision, going back to his digital archive and retrieving files and images and video clips that might pertain to that information, and that decision at hand? If so, I think that his ability to decide, in time, might become impaired through the use of his life-blog tools, his digital memory. But I don't know whether he is actually using it for that purpose or not. Perhaps he is only using and creating the life-blog archive, the perfect comprehensive memory of what he encounters, just because he wants to find out whether it is possible to do it.

Nora:   Where do you think we are headed, Viktor? Do you think we will see a return to forgetting?

Viktor:   I hope. I hope we see a return to forgetting, because I think this will enable us to not be overwhelmed by the digital memory that we create, by the digital past that is being recorded and kept in storage. I hope that we'll be able to overcome this hurdle, like we have overcome so many other hurdles before. I remain very optimistic about that, but I do think it's going to be a hurdle, and a significant one. And I do think we need to take the challenges and the threats of digital comprehensive memory very seriously.