A Computer Scientist Makes the Case for Speculative Fiction

“PREDICTION IS VERY difficult, especially if it’s about the future,” said Physics Nobel Laureate Niels Bohr. Bohr was presumably talking about the vagaries of quantum mechanical subatomic life, but the statement holds true at other scales too. Predicting the future is tough, and any good scientist knows enough to hedge his or her bets. That’s what error bars are all about. It’s why science usually proceeds methodically: hypotheses are formulated, experiments conducted, observations collated, and data evaluated. Almost by definition, science is intolerant of fanciful speculation — and especially so in the current climate, when external funding from almost all sources requires that pure research have near-term applications. An acutely myopic memo issued by the White House on August 17 of this year states that “agencies should give priority to funding basic and early-stage applied research that, supplemented by private sector financing of later-stage R&D, can result in the development of transformative commercial products and services.” In short, the aspiring scientist had better be able to argue for a short path from dreams to deliverables.

Unfortunately, just when scientists are being encouraged to narrow their sights to the next bankable application, we as a society most desperately need true visionaries — and not, mind you, self-branded futurists — willing to peer far over the scientific horizon. As it turns out, history itself suggests we turn not to scientists but to artists, especially novelists and filmmakers. Unimpeded by error bars and immune to the stakes that motivate futurists, they may be our best guides to the science-inflected possibilities of the future, precisely because they have the freedom to chart a possible future that many a bench scientist can’t or won’t. Physicist Leo Szilard, credited by some as the father of the atomic bomb, once wrote to an industrialist friend that when it came to predictions about the applications of new understandings of the atom, “[T]he forecast of the writers may prove to be more accurate than the forecast of the scientists. The physicists have conclusive arguments as to why we cannot create at present new sources of energy for industrial purposes; I am not so sure whether they do not miss the point.”

Szilard had in mind H. G. Wells’s 1914 novel The World Set Free. Often cited as the earliest source for the idea of harnessing atomic energy for weapons development, it sensitized Szilard and educated Winston Churchill to the martial possibilities of the atom, helping spur Churchill’s initiation of a British A-bomb effort. Coupling whiz-bang adventure tale and popular science, it was a form of literature particularly well suited to turn-of-the-20th-century England, when exploration (and its dark twin: colonialism) and scientific achievement were broadly celebrated and communicated. In his fiction, Wells also foresaw television, the VCR, air warfare, and even genetic engineering. His prescience was among the reasons Churchill considered him a close advisor. More broadly, Wells’s literary talents introduced a generation and more to the implications of scientific achievement.

Today we call works like those of Wells “speculative fiction” or “near-future fiction.” Given the pervasiveness of technology in our daily lives, not to mention a general jittery uncertainty about the state of world, it’s hardly surprising that bookstores and theaters are again overflowing with creative work exploring the ramifications of our newest scientific and technological advances. Much of speculative fiction is dark, ironic, and satirical, situating worlds in an uncomfortable liminal space between the real and a dark near-real. The scariest of them feel the most probable given the world as it is. We’d be well served to pay attention to their hypothetical “what if” worlds.

Wells referred to writings like The World Set Free as “fantasias of possibility” in that each “takes some great creative tendency, or group of tendencies, and develops its possible consequences in the future.” His “creative tendencies” might be interpreted as technological innovations. Creative and scary magical realism — great stories, in other words — can happen when you mix those innovations with the tendency of groups to rapidly evolve new social norms.

Wells effectively asked both “Can we do this?” and “Should we do this?” He imagined a world with atomic weapons and, concomitantly, a global arms race and the economic, social, and political problems it might induce. As it turned out, Szilard also asked both these questions, helping to start the research that resulted in the atomic bomb while also taking on a pioneering role in advocating for nuclear disarmament. To be sure, the ethical and societal implications around the invention of “the bomb” paled with the realities of World War II, so perhaps no amount of inspiring literature could have — or should have — derailed its production. But works like Wells’s may very well have played a role in shaping Szilard’s thoughts — and the thoughts of others pondering its downstream societal effects.

As a computer science and mathematics professor, I had hypotheticals in mind when I last taught a class on machine learning at Dartmouth College. My students were a group of mathematically adept juniors and seniors, all of them standing a good chance of pursuing careers in “next-gen” data-driven technologies. The potential long-range societal effects of such creations are rarely — if ever — discussed before deployment, only post-hoc at best. Contemporary and even classic speculative fiction does in a sense frontload the discussion, which is why I thought to use this literature to sensitize my students.

Machine learning, or “ML” as we often call it, encompasses a wide range of techniques and technologies that do in fact revolve around prediction — Bohr’s admonition against predicting the future be damned, sort of. Algorithms are developed that generally, with a quantifiable degree of confidence, use the past to make guesses about the future. Applications range from the seemingly anodyne (pushing purchases) to overtly life-altering (predicting terrorist activity). ML enables driverless cars and workerless factories. It often depends on the availability of relevant “big data” in order to take advantage of large numbers of exemplars. This kind of methodology works less well — if at all — when deployed on the scale of nations and societies. So, yes, Bohr is right in saying prediction is difficult — I think — when working at the scale of the atom or how people behave en masse, even if he may be wrong when it comes to the individual and what book she is likely to buy next on Amazon.

As befitting an advanced class in applied mathematics, my students were asked to complete a number of technical assignments, including an intensive large-scale project, which for many involved the analysis of some kind of “big data.” But what was slightly different was the class’s reading/writing arm: students were also required to read and then respond to one of several books critiquing “the digital” including the two best-selling speculative fictions, Dave Eggers’s The Circle and Gary Shteyngart’s Super Sad True Love Story. My hope was that the real-time integration of some speculative critique might engender in my students an ethos of what I like to call “thoughtful technology deployment.”

As an educator, mathematician, computer scientist, school board member, and parent of young children, I have been dismayed by how little reflection drives the development and deployment of technologies — and continually struck by the “hegemony of the new.” Technologies are often introduced into environments, especially educational ones, from fear of being left behind rather than from a principled data-driven belief in anticipated benefits. It seems to me that reading a little speculative fiction might counter our fear of being left behind with some modicum of caution about moving ahead. Grounded conversations — even if grounded in fanciful speculation — could very well modulate how we introduce “the new” into our lives and classrooms.

Take the recent epidemic of shaming: social media now enables finger-pointing on the scale of millions. This phenomenon is most masterfully illustrated in Eggers’s The Circle, which follows the rise and fall of Mae, an employee of a Facebook-Google mashup called The Circle. Today, Facebook claims to have over two billion active users a month, which amounts to more than a quarter of the planet sharing their lives with “friends.” Meanwhile, passive and aggressive surveillance is increasing in the workplace, on the street, and at home, chipping away at our understandings of privacy and personal space. Eggers takes up Wells’s gauntlet and “suppose[s] these forces to go on” to the natural endpoint of complete transparency. I advise you to forget the forgettable movie version starring Tom Hanks and Emma Watson and go directly to the book to see what happens when your every action is fodder to be “liked” (or not) by pretty much everyone. Not hilarity but mob rule ensues. Given current trends, is Eggers’s vision of the future already here? Are there enforceable ways to modulate transparency? Who should be in charge? How does the binary nature of “like” contribute to mob rule? These are the kinds of questions that can be concretely discussed when real events are amplified in fictional narratives — and, arguably, they should be discussed while the technologies are being created.

The dystopic downsides of transparency plus data also drives Gary Shteyngart’s Super Sad True Love Story. This satiric novel follows the misadventures of Lenny Abramov, a sad-sack upper-level employee of a company called Staatling-Wapachung Corporation that focuses on Indefinite Life Extension, which may or may not actually have the secret to indefinitely extending your life. Like the rest of the world, Lenny lives tethered to his smartphone-on-steroids “äppärät” in a thoroughly militarized United States deeply in debt to China. The reader learns the lay of the land in part through Lenny’s diary, one of many nods — including a good deal of stylized NewSpeak (“LandO’LakesGMFordCredit”!, “AlliedWasteCVSCitigroupCredit”!) — to Orwell’s 1984. But where 1984 is relentlessly gray and leaden, enveloped in a shroud of Stalinist totalitarianism, Super Sad True Love Story has the vibe of ADHD or an out-of-control cartoonish video game, appropriate to the generation it sends up.

Shteyngart exploits Wellsian tendencies in our current world, from the gradual encroachment of Homeland Security on our daily existence to the ubiquitous monitoring and broadcasting of our personal information via our smartphones and Fitbits. In a classroom of future developers and policymakers, it would make sense to have conversations now rather than in the near-future USA around how little it would take to divert your digital stream from the cloud to a store or person you pass in the street, and what effects this would have.

Were I to teach my class again — and I will — I would put a third book on the list: another recent dystopian best seller, Station Eleven by Emily St. John Mandel. Exploring our desire to preserve culture as a connective social glue (among many other themes), its starting point is a fast-moving worldwide epidemic. Word of a virus makes its way onto the news, but for all the talk of the great pandemics of the past, all of us know that in today’s world “the flu” comes and goes — until it doesn’t. The speed of spread and the efficiency of the virus precipitate a crash of all communications infrastructure. The few who survive the virus do so for contingent reasons.

Whereas Eggers and Shteyngart make their points through exaggeration, Mandel makes hers through reduction. Nostalgia is embodied in a minimally preserved culture — manifested in hunger for any artifact that existed before the collapse. Scraps of newspapers become cultural relics. A remnant of a graphic novel is a holy document of a new religion. An accumulation of odd objects from the days before the epidemic turn into a “Museum of Civilization” in an airport that avoided the deadly virus because it happened to turn away planes with infected passengers. The museum becomes a mecca for those left alive who want to rebuild the world by turning the lights back on. A roving band of players, “The Traveling Symphony,” puts on Shakespeare for anyone who will agree to watch, but kills when it needs to. Mandel has a light touch in portraying the likely speed of societal collapse, which, she suggests, comes with the ease of global travel, dependence on the internet and power grid, and the fragility of a culture kept in the cloud and the Internet Archive. So, in short — what are we left with when the lights go out? How much do culture and social norms depend on keeping the lights on?

While not overtly technological in theme, I can’t help also mentioning Octavia Butler’s 1990s “Earthseed” series. It, too, belongs in my future course. Unlike Mandel’s turn-on-a-dime dystopia, Butler’s disaster scenario portrays a near-future dystopia that unfolds like a slow train wreck. Placed in the 2020s and 2030s, her future can seem so close that you can almost touch it, the tendencies of today’s United States advanced ever so slightly. The rich live in armed encampments and the poor in versions of tribal tent cities. The latter coexist in a state of ever-simmering tribal warfare, amplified by attacking gangs of drug addicts, high and hopeless, eager to burn, rape, and kill. Stores still function, albeit under armed guard. Police remain distinct from soldiers in armies, and so people still look for justice, if warily. Guns are rife in this open-carry world, and families still wrestle with sex education and gun control. But the country is collapsing, with states instituting their own local immigration policy as the oval office dissolves. Slavery has returned as a “solution” to illegal immigration and debt. There is even a divisive, misogynist demagogue presidential candidate whose platform focuses on a return to simpler Christian times while urging followers to “help us to make America great again!” I kid you not. Butler’s prose is clean and impressionistic, allowing you to fill in the details of how we got there. In her post-climate-change world, the coast has been pulled inland and clean water and food are scarce, although not unattainable. Fiction? Sure — although if you read Zeitoun, also by Eggers, his nonfiction account of the harrowing post-Katrina experience of a Muslim-American family, and follow some of the stories of post-Harvey/Irma/Maria events, Butler may appear more visionary than we would like.

The Circle, Super Sad True Love Story, and Station Eleven are relatively new works, but the current political situation has by many accounts brought renewed interest to the works of Wells as well as to other classics, like George Orwell’s 1984 and Aldous Huxley’s Brave New World. As of this writing, Margaret Atwood’s 1985 novel The Handmaid’s Tale is in its 17th week on Amazon’s top 20 most-read list (surely helped by the new Hulu series based on the book).

My hope is that discussing these texts in AI classrooms will prod a few nascent developers to think differently about information centralization and ownership. Of course I cannot yet know whether my little classroom exercise will have downstream effects, but my students’ responses suggest that the readings at least gave them pause. “How connected do we really want to be?” ask my students after reading Shteyngart. They don’t have answers at this point, but, coupled with the experiential learning of data-intensive projects, such required reading might make for a breed of more thoughtful technologists. And, going straight to the source: What if big tech companies had all of their employees read a book or two in this genre each year, and then asked whether this is the future they are building? And the follow-up question: Is this the future we want? And better still: Is it the future everybody else wants? Some well-chosen reading might help us hit the pause button once in a while. As George Santayana might have said, those who fail to learn from the future may very well be doomed to produce it.


Dan Rockmore is associate dean for the Sciences and the William H. Neukom 1964 Distinguished Professor Computational Science at Dartmouth College, the director of the Neukom Institute for Computational Science, and professor of mathematics and computer science. He is editor of the recently released volume, What are the Arts and Sciences? A Guide for the Curious, from UPNE and is a member of the external faculty of The Santa Fe Institute.