recent, Science / Tech, Tech

The Fallacy of Techno-Optimism

When the telegraph revolutionized communication in the 1850s, Henry David Thoreau was scornful. “We are in a great haste to construct a magnetic telegraph from Maine to Texas,” he wrote, “but Maine and Texas, it may be, have nothing important to communicate.” Thoreau, it turned out, was wrong, but his skepticism of an invention that would ultimately prove beneficial has been a common response to change, both in the past and the present.

A fascinating collection of these responses exists on a podcast called Pessimists Archive, the episodes of which catalogue the unjustified fear that greeted innovations like the telephone (blamed at the time for degrading social life), the bicycle (blamed for various health maladies), and the novel (blamed for corrupting youth). The podcast presents these overreactions as cautionary tales for those who would question the fruits of Silicon Valley in the present, and its message has famous fans: Steven Pinker tweeted that the podcast is “invaluable in putting today’s tech moral panics in historical perspective.”

Indeed, perspective will always be valuable. But I suspect that Pessimists Archive and its fans are driving at a more ambitious point: that because past fears of change were unjustified, so are present fears. If the telephone and the bicycle and the novel worked out, so too will AI, driverless cars, and automation. The podcast’s producers make that very case in their inaugural episode, which intones a manifesto:

We all know how the phone, the bicycle, and the novel turned out. We love them. The fears surrounding them were foolish. So it’s worth asking, why do we repeat ourselves like this? Why do we always say “no, this time it’s different, this time we’re in danger.” Why can’t we trust our own history?

This sentiment embodies a popular kind of techno-futurist confidence that newness is progress, which must triumph over the hidebound conservatism that impedes it. It believes that the arc of history bends toward disruption, and that our own worries about transformative technologies will one day appear as irredeemably silly as Thoreau’s rejection of the telegraph. With that, the true value of Pessimists Archive emerges: exposing technological optimism as little more than a fallacy, almost stupefying in its wrongness.

“Trusting our own history” means relying on what happened in the past to tell us what will happen in the future. That’s a recipe for getting it wrong. When forecasting the future, perhaps the only thing that can be trusted is the emergence of unprecedented, unpredictable events that violate past trends. On the eve of the 2008 subprime mortgage crisis, none of the models used to forecast house prices accounted for the possibility of a price collapse—for the simple reason that no such collapse had ever happened. Price data was historical, and extrapolating that history into the future rendered us blind to the possibility that something ahistorical might happen. Or as Pessimists Archive unwittingly put it: the possibility that “this time it’s different.”

But techno-optimists take the exercise a step further, by using data about one thing to forecast the future of an entirely different thing. This moves us from the flawed to the absurd. For instance, when techno-optimists compare anxiety over driverless cars to the protests of the horse-and-buggy industry over the automobile, they ignore the ways that driverless cars implicate fundamentally different problems than did automobiles. Driverless cars stand ready to collect immense amounts of personal data about the habits of their passengers, and their networked structure creates serious national security risks. What could the successful debut of the automobile in the early 20th century possibly tell us about that?

The political philosopher Gerald Gaus argues that the less precedent there is for some practice working, the less reason there is to prefer it. In the case of new technologies that implicate brand-new problems, the data we have about the success of past technologies is simply irrelevant. History gives us no reason to prefer a world in which, for example, most manual work is automated. It’s never happened before.

We draw spurious historical analogies precisely because it’s easy. If we can say that Change A is just like Change B, which went fine, it allows us to avoid grappling with the actual qualities of Change A. Argument by analogy displaces argument on the merits. It’s far more convenient to assert that past change was good, present change is just like past change, therefore present change is good too.

Unfortunately, by any objective measure, most new things are bad. People are positively brimming with awful ideas. Ninety percent of startups and 70 percent of small businesses fail. Just 56 percent of patent applications are granted, and over 90 percent of those patents never make any money. Each year, 30,000 new consumer products are brought to market, and 95 percent of them fail. Those innovations that do succeed tend to be the result of an iterative process of trial-and-error involving scores of bad ideas that lead to a single good one, which finally triumphs. Even evolution itself follows this pattern: the vast majority of genetic mutations confer no advantage or are actively harmful. Skepticism towards new ideas turns out to be remarkably well-warranted.

The need for skepticism towards change is just as great when the innovation is social or political. For generations, many progressives embraced Marxism and thought its triumph inevitable. Future generations would view us as foolish for resisting it—just like Thoreau and the telegraph. But it turned out that Marxism was a terrible idea, and resisting it an excellent one. It had that in common with virtually every other utopian ideal in the history of social thought. Humans struggle to identify where precisely the arc of history is pointing.

Techno-optimists would likely prefer to put aside failed products and ideologies and consider instead those innovations that have already proven successful. We’re talking about the iPhone, after all. Is popular adoption of an innovation reason enough to suspend skepticism? No—we turn out to be quite bad at predicting the full impact of even our most successful ideas. Adding lead to gasoline made automobile transportation more efficient, but it caused widespread brain damage and may have been responsible for the 20th century crime boom. Freon in refrigerators punched a hole in the ozone and had to be banned by international compact. Fossil fuels, one of the most successful product innovations in history, are experiencing what might politely be called a re-evaluation.

Another massively successful innovation undergoing a re-evaluation of its own is the internet itself. Optimists promised emancipation: knowledge would be democratized and civil discourse would flourish. Now, we understand that the internet is also a highly effective system of control. Incentives to commodify personal information have resulted in more and more of our daily lives becoming subject to data collection, transforming our economy into a surveillance ecosystem. This renders our conduct “visible” to states, which can punish us for it—as China is doing now through its dystopian “social credit” system. The whole thing could turn out to be a terrible mistake—we don’t know, because we’ve never had to solve this problem before. The fact that we previously solved the problem of the telegraph is irrelevant. One could probably fill a podcast—call it the “Optimists’ Archive”—with inappropriately rosy predictions about the wonders of new technology.

We are engaged in a giant social experiment. For 99 percent of the time humans have lived in settled societies, life in each generation was essentially like life in the generation before. Stasis, not change, was the rule. Now, for the first time, we live differently, and the gap between the generations grows wider as the pace of change grows faster. Can this continue indefinitely? We have no precedent for that working. Analogies to history are analogies to nothing at all. We might as well analogize the driverless car to the hand-ax.

Instead of empty analogies, the only way to survive change is to have a vigorous debate about the merits of our new ideas—precisely the kind of debate that techno-optimists want to foreclose by appealing to history. We might ask instead: what does this new thing do to us? Do we understand enough to answer that question? If not, on what basis does our confidence rest? Debate on the merits is essential to distinguishing good ideas from bad ones. And for that, you need the people that techno-optimists most loathe: conservatives.

Liberals and conservatives don’t just vote for different parties—they are different kinds of people. They differ psychologically in ways that are consistent across geography and culture. For example, liberals measure higher on traits like “openness to new experience” and seek out novelty. Conservatives prefer order and predictability. Their attachment to the status quo is an impediment to the re-ordering of society around new technology. Meanwhile, the technologists of Silicon Valley, while suspicious of government regulation, are still some of the most liberal people in the country. Not all liberals are techno-optimists, but virtually all techno-optimists are liberals.

A debate on the merits of change between these two psychological profiles helps guarantee that change benefits society instead of ruining it. Conservatives act as gatekeepers enforcing quality control on the ideas of progressives, ultimately letting in the good ones (like democracy) and keeping out the bad ones (like Marxism). Conservatives have often been wrong in their opposition to good ideas, but the need to win over a critical mass of them ensures that only the best-supported changes are implemented. Curiously, when the change in question is technological rather than social, this debate process is neutered. Instead, we get “inevitablism”—the insistence that opposition to technological change is illegitimate because it will happen regardless of what we want, as if we were captives of history instead of its shapers.

This attitude is all the more bizarre because it hasn’t always been true. When nuclear technology threatened Armageddon, we did what was necessary to limit it to the best of our ability. It may be that AI or automation causes social Armageddon. No one really knows—but if the pessimists are right, will we still have it in us to respond accordingly? It seems like the command of the optimists to lay back and “trust our history” has the upper hand.

Conservative critics of change have often been comically wrong—just like optimists. That’s ultimately not so interesting, because humans can’t predict the future. More interesting have been the times when the predictions of the critics actually came true. The Luddites were skilled artisans who feared that industrial technology would destroy their way of life and replace their high-status craft with factory drudgery. They were right. Twentieth century moralists feared that the automobile would facilitate dating and casual sex. They were right. They erred not in their predictions, but in their belief that the predicted effects were incompatible with a good society. That requires us to have a debate on the merits—about the meaning of the good life.

We may wake up one morning and find that our most successful innovations are actually incompatible with the good life. By then it might be too late to do anything about it. We are no longer ship captains charting a course to a destination, we are passengers in driverless cars, riding the arc of history. Where it will take us is anyone’s guess.

 

Nicholas Phillips is a law student and writer based in New York. He is a research associate at Heterodox Academy. Follow him on Twitter @nicholas_c_p