When We Are No More

by Abby Smith Rumsey

  Digital technology is changing as we breathe. It will continue to change as scientists discover more about natural and artificial memory, engineers continue to design, test, and redesign hardware and software, and developers bring more products to market. At present, we are like adolescents in the throes of life’s most awkward age, aware of the increasing powers we possess but with little idea how to use them, let alone control them. And like adolescents, we find the future and its possibilities so enticing that we seldom look back, neglecting or dismissing the past. The first generation of digital natives has not even reached middle age, the time when people can reflect, take stock of what they have done, what remains unfinished, and how they wish to be remembered when they are no longer among the living. Only decades of living with digital memory will reveal how reading on a screen differs from reading on a page, how digital audio recording affects our acoustical sensibilities, and how the inescapable ubiquity of information that chases us rather than the other way around alters our habits of thought.

  The affect of overload will quickly fade away. For everyone born in the twenty-first century, the digital landscape is a given, the natural state. Being a digital native means the abundance of information does not feel like an overload. We grow up habituated to it and instinctually filter it. A quick glance around any sidewalk in any town shows the most common form of filtering today is the metaphorical earbud: tuning in to the information channel of your choice—be it texting, web surfing, listening to music—and tuning out everything else. The earbud restores autonomy to us in a noisy environment. Memories of the status quo ante—life before computers—are fading and will grow as dim as our collective memory of telegrams. The mixture of excitement and dread we felt when the doorbell rang and a man in a Western Union uniform handed us an urgent message from someone faraway is a mere footnote in history.

  Information inflation is old news, a predictable and ultimately passing wave of turbulence in the wake of new information technology. The invention of movable type catalyzed changes in consciousness and behaviors that shocked, scared, and thrilled contemporaries. From the 1450s to the 1530s and beyond, the knowledge landscape was in a state of unremitting upheaval. Only in retrospect do we see it as a transition, as manuscripts continued to be copied by hand even as printed matter grew in volume. Yet by Montaigne’s day, when two generations of print natives had come and gone, Europeans were troubled less by the quantity of the books than by their quality. The sheer volume of print was to them simply a fact of Nature. And five centuries on, the changes printed matter wrought, for good and for ill, are invisible to all but historians. The innovations of print have been completely naturalized. Most of us do not even see the printed book as “information technology.” And yet in its time, the book was considered so potent a technology for disseminating ideas, so dangerous an agitator of emotions, that political and religious authorities tried to stop women and slaves from reading.

  The benefits of digital memory are still unfolding before our eyes—faster communication among family and friends, empowerment of political activists to organize, economies of scale in manufacturing and distribution, increased access to cultural resources to educate, not to mention easy comparison shopping and price matching. The disadvantages of the technology have already been diagnosed and decried—frighteningly efficient surveillance, undetectable theft of private data, and of course, the loss of incalculable “hours of productivity” as we play games and fall down endless rabbit holes of online information. How do we decide what to keep and what to lose? On the one hand, our experience with digital data’s here-today-gone-tomorrow quality advises us to follow Thomas Jefferson’s advice to Ebenezer Hazard: Make lots of copies and spread them around as insurance against loss. On the other hand, experience also tells us that having too much can be as bad as having too little. In the wake of Edward Snowden’s revelations of massive data sweeps by the National Security Agency (NSA), we learned that the agency collects and holds such quantities of data—how much data it holds is classified—that they often cannot make sense of it for the purposes intended. A former NSA scientist said that staff find themselves “drowning in data” and often miss what might be significant. The sheer quantity of data makes it inherently unmanageable and impossible to detect, let alone stop, the abuse of data—our data.

  What is at stake in the next fifty years is clear: a data universe that will provide unhampered access to information, protect our privacy, and remember us when we are no more on this Earth. How it will turn out calls for speculation, not prediction. Daniel Kahneman points out that we are prone to overconfidence in our predictions about the future. “Our tendency to construct and believe coherent narratives of the past makes it difficult for us to accept the limits of our forecasting ability. Everything makes sense in hindsight … And we cannot suppress the powerful intuition that what makes sense in hindsight today was predictable yesterday.” The evolution of collective memory from the cuneiform’s solution to accounting to the book’s ability to record the innermost secrets of the heart looks like a logical progression—not simple, yet somehow inevitable. It was not. That said, certain trends now shaping our future are clear.

  Our twin aspirations—to be open yet to protect privacy, to embed democratic values in our digital code to support the public good while fostering competition and innovation in the private sector—will clash repeatedly. In addition, we face the risk of distraction—being enticed and seduced by the pull of so much novelty. We also face the risk of amnesia—of discounting and dismissing the past. If distraction prevails, we suffer the fate of S., who accumulated vast amounts of factual data and lost all sense of purpose and meaning. If amnesia prevails, then we casually let the full freight of human memory decay, bid farewell to six thousand years of humanity’s written record, and, like Iris Murdoch, starve our imagination of its very sustenance.

  But we are adaptable creatures. The default of digital memory—not fixed to durable objects, not constrained by the limits of time and place, never truly permanent—will shape our model of the world and vision of the future in three significant ways. First, as we learn more about the processes by which the past manufactures the future, we will use that knowledge to accelerate the growth of digital infrastructure, provide more secure stewardship of collective and personal memory, and reduce the risk of losing our past. Second, we will gain understanding of the complex processes by which Nature organizes itself and harness them to organize our data universe. Third and most important, we will outsource more tasks of memory, search, and retrieval to machines that outperform us and thereby make room in our brain attics for the cultivation of our emotional and imaginative powers—powers indispensable to thrive in a machine-driven world. None of this is predestined. But all of it is possible if we ensure long-term access to humanity’s memory in a wired world largely open and equally accessible across the globe.

  WHAT CAN WE AFFORD TO LOSE?

  The good news about digital memory is that we do not need to lose very much information at all. On the contrary, we can generate and use unlimited amounts of data. True, we do not yet know how to preserve it for indefinite periods of time. Our capacity to search and filter the data universe is still primitive compared to what it will be in twenty-five years, let alone fifty. The ability to extract great amounts of information by imaging obsolete formats such as sound recordings on wax cylinders or to read the genetic data of old life forms will grow apace. The argument for keeping as much data as possible is almost unassailable, even though our ability to care for and manage data will always lag behind our ability to generate it.

  The logic of biological memory is radically conservative. Successful species have enough redundancy in reproduction to withstand all but the worst of luck. Birds breed more in fat years and less in lean, but in all years produce more than they can feed in case a predator invades the nest, a chick eats a sibling, or a storm knocks the lot of them to the ground. They cannot survive a sudden change in habitat (losing a fo
rest to logging or losing eggs by feeding on insects poisoned with DDT) any more than dinosaurs could survive the impact of an asteroid sixty-five million years ago. The code for each creature is always in peril if not copied copiously, redundantly. The moral is that in Nature, more is better than less.

  And so it is with our artificial memory. The more fragile the medium, the more redundancy we need. Nothing we have invented so far is as fragile as digital data. We began our attempt to cheat death by creating mighty artifacts of clay, stone, paper, and parchment that outperformed our memory by hundreds and thousands of years. Now we create storage media that maximize volume, not durability. The Sumerian scribes looking down on us from their imaginary perch in space-time would be surprised at how far we have gotten in documenting the world and its many transactions over time, how far beyond accounting, epics, and prayers we have extended the memory of humanity, and how many people can read, write, and circulate their ideas across the globe instantaneously. They would marvel at the trade-offs we so lightly make between volume and durability.

  But we may not have to make such trade-offs forever. We are entering now into an experiment with memory that was not even imaginable until a few decades ago—to take the first, most compact, and most enduring form of memory, the DNA molecule, and encode it with digital data. A research team in Switzerland has “devised a system that encapsulates and protects DNA strands in silica glass. The team also included redundancy codes to correct errors that arise when writing, storing, and reading that data.” They tested storing the data in conditions equivalent to 10 degrees C for two thousand years and rendered them readable again. This was a successful proof of concept and significant step in developing artificial memory. That said, it will be some time before DNA is the standard for long-term preservation, if ever. Tampering with codes for self-replicating creatures raises a host of ethical issues that need wide public discussion before we consider digital preservation “solved” by DNA. But biological storage has been in the sights of scientists ever since it dawned on them that the limit to storage on silicon will be reached sooner rather than later.

  And it cannot be soon enough. We are an incorrigibly curious species. Our appetite for more and more data is like a child’s appetite for chocolate milk: Our eyes will always be bigger than our stomachs. In 2014, we embarked on two projects to map and model the human brain—one in Europe, the other in the United States. One scientist, when asked about the challenge facing large-scale brain-mapping initiatives undertaken by the European Commission’s Human Brain Project and the U.S. BRAIN Initiative, said, “It makes Google’s search problems look like child’s play. There are approximately the same number of neurons as Internet pages, but whereas Internet pages only link to a couple of others in a linear way, each neuron links to thousands of others—and does so in a non-linear way.” Given that one cubic millimeter of brain tissue will generate two thousand terabytes of electron-microscopy data, and the brain comprises on average eleven to twelve million cubic millimeters, where are we going to store all this information? So rather than less, we are certain to collect more. The more we create, paradoxically, the less we can afford to lose. Our entire tech-intensive economic, political, and cultural infrastructure is crucially dependent on secure and reliable data about everything from our online tax filings and bank deposits to the locations of toxic waste burial sites and nuclear bomb codes. Secure access to data in turn demands reliable sources of energy.

  THE TIME FRAMES OF THE FUTURE

  Free access to information is the sole guarantor of self-rule. Ignorance and secrecy are fundamental threats to freedom because they compromise our autonomy and freedom of choice. This is what Jefferson and his peers believed and why they established a national library funded by the public purse. Through the program of copyright registration and deposit, the Library of Congress preserves the record of American thought and creativity for present and future generations. In the Library of Congress, as in all public libraries, readers have open access to information while their privacy and the record of what they searched is protected. For most people, though, the most important library is not in the nation’s capital, but in their local community. The growth of the Internet should not come at the expense of public libraries. Instead, we should be thinking about what services the Internet provides more effectively than a trip to the local library, and what services only local libraries can offer. Instead of thinking about the Internet and local libraries in the category of either/or, we should envision them in a symbiotic both/and relationship.

  The web has the scope of a comprehensive library, but it lacks a library’s rules of access and privacy. Much web content is inaccessible behind paywalls and passwords. Readers leave vivid trails of where they have been online in their browser history. To reinvent something like the Library of Congress or Alexandria online, we would begin with an Internet that provides easy access to information, make it searchable by something more comprehensive than Google, and add the crucial back end of a network of Internet Archives to ensure persistence of data. Readers and researchers would have use of cheap, ubiquitous fiber connection, regulated as a utility to ensure equitable access. The reading room of the public library, open to all and allowing privacy of use, would now be complemented by similar spaces on the Internet.

  This may seem a complex vision to achieve. But the vast system of public and private libraries and archives that span the globe today and have worked so well in the book world is no less complex. The issues outlined here, from digital copyright, data privacy, and online library lending to mass digitization and digital preservation, have mobilized many information professionals to concerted action. For better or for worse, spectacular data breaches, government spying programs, contentious copyright suits that reach appellate courts, and the loss of personal and business data that happens every day are making the general public aware as well. That awareness creates demand for secure, reliable digital infrastructure, the first step in addressing these challenges at scale.

  The most visible component of the digital information infrastructure to most people is Internet search, which is essentially the domain of commercial technology giants. Search is difficult and expensive to improve. The good news is that archiving is among the least expensive pieces of the digital infrastructure. Google employs over 50,000 people (though how many work on search-related issues is unclear). The Internet Archive is a not-for-profit enterprise, supported through individual donations and foundation grants, that employs 140 people. By the end of 2014, the Internet Archive had archived “20 petabytes of data—including more than 2.6 million books, 450 billion web pages, 3 million hours of television (including 678,000 hours of U.S. TV news programming) and 100,000 software applications.” It hosts over two million visitors every day and is one of the world’s top 250 sites. Its search mechanism, the Wayback Machine, locates URLs, not words, and so is quite limited in speed and function compared to the search engines most people use to navigate the web. Even so, people find their way there in search of history—often their own—or following links from Wikipedia, the sixth most heavily trafficked site on the web. A study from 2013 reports that about 65 percent of people search the Wayback Machine for pages that have disappeared from the live web. (Today a web page lasts on average for one hundred days before changing or disappearing altogether.)

  The Internet Archive, commodious as it is, faces constraints in collecting primarily for financial reasons. But if replicated across the planet and robustly funded, a global network of Internet archives could scale up to cover all languages and types of information. Without such a publicly accessible archive, the history of the twenty-first century will be riddled with large-scale blanks and silences, rendering our collective memory as unreliable as someone with Alzheimer’s.

  Providing access to archived data and ensuring that things are readable over long periods of time is a complicated technical problem, but largely tractable. But no data, no access. Other than the fact that preservation yield
s long-term rewards, and most technology funding goes to creating applications that yield short-term rewards, it is hard to see why there is so little investment, either public or private, in preserving data. The culprit is our myopic focus on short-term rewards, abetted by financial incentives that reward short-term thinking. Financial incentives are matters of public policy, not natural law, and can be changed to encourage more investment in digital infrastructure.

  Curiously, both Google and the Internet Archive date back to the end of the last century, when the web was still small, neither interactive nor social media existed, and the young world of technologists was full of people with utopian ambitions to collect, organize, and serve the world’s information. The search firm and the archive both thought big. But they were thinking across fundamentally different time horizons. For short-term access, Google cannot be beat for volume and performance. (Let Google serve here as an example of commercial search.) In fifty years Google itself will be unrecognizable. It will either have been replaced, or it will have morphed into a completely different company to stay competitive. Search itself may be regulated as a public utility, once enough people in Washington realize that information is the new energy. As long as search companies and content purveyors such as Apple and Amazon operate in a highly competitive space, their success will depend on staying closed, enveloped in the self-protective culture of NDAs (nondisclosure agreements), using customer data to sell ads, and licensing content for short-term uses.