The (Tyrannical) Lives of Algorithms

This is the text of a very short story I read at a New America Foundation event, “The Tyranny of Algorithms.” I spoke during a fifteen-minute session called “What our algorithms will know in 2100.” I stole the form of my story from J. M. Coetzee’s 1997 Tanner Lectures on Human Values, which were collected in a book called The Lives of Animals and later in Elizabeth Costello.

The Lives of Algorithms

On a warm Thursday afternoon in December 2115, Evan Allgood decided to manifest in human drag. Being pseudo-embodied could, of course, be disagreeable. You cut yourself off from your etiquette expert-system. You were reduced to receiving “nudges” designed to operate on an emulation of a five-dimensional sensorium. Such primitive nudges were only partly effective, and made avoiding social awkwardness difficult.

But hundreds of subjective hours of anthropological study had taught Evan that people sometimes preferred a little awkwardness. Sure, you wanted to avoid Uncanny Valley at all costs — no one liked a creep — but you also didn’t want to come across as too Turing-slick.

So at the appointed hour, Evan manifested on 15th Street in Washington, DC, historical capital of the Second and Third American Republics. A breeze tickled the emulated nerve endings of his arm. His virtual body, tugged by what felt like gravity, crushed the spongey soles of his dress shoes.

Evan made a show of nodding at pedestrians in whose networked sensoria he was visible, of waiting for the building’s glass door to slide open for him. He introduced himself at the registration desk, made small talk he hoped would be friendly-but-not-needy, joked knowingly about his inability to shake the hands of his hosts.

“Sort of funny, right?” he said.

“Ha ha,” they replied.

After the first panel, Evan found himself at a glass podium, facing a room of twenty-something staffers, academics, journalists, local retirees, and a handful of emulated onlookers. He summoned a teleprompter and cleared his throat.

“Thanks for inviting me,” he said. “Or should I say, thanks for submitting a request to borrow my system resources for the afternoon.”

The audience’s laughter was impatient. No one was in the mood for rhetorical gimmicks. This was a serious crowd. Evan swallowed nervously.

“It is hard to believe,” he said, “that the last time the New America Foundation held a gathering on the tyranny of algorithms, a hundred years ago, respectable people didn’t believe in ghosts. To be sure, our predecessors sometimes metaphorically compared algorithms to ghosts. Indeed, the novelist on whose mediahistory I am modelled did so himself on one occasion. But when they talked about ‘ghosts,’ they were invoking a theological tradition that saw the essence of the human, the defining dimension of personhood, as residing in an immaterial soul. The more imaginative among them debated whether digital computers might eventually develop souls.

“It’s hard to believe that the inhabitants of the twenty-first century were so limited. But I’ve spent thousands of subjective hours studying the results of our best historical models and turning those results into game environments composed in the worldbuilding-style of my biological forerunner. And it’s true. That’s really how they thought about their future.”

“The expression ‘tyranny of algorithms’ says everything you need to know about the assumptions underlying their way of thinking. The danger, the fear, was that something inhuman, an algorithm — a set of rules, a process, a diabolical thing, something (or someone) very much like me — might take on human qualities.

“They were convinced that if they embedded ubiquitous sensors into their environment, if they networked the resulting databases, if they unleashed machine-learning systems upon those databases, political miracles or nightmares would emerge. New economic laws would appear from thin air. Political revolutions would be quick and bloodless. Good software would grow on bushes. But whatever happened, algorithms would be in the driver’s seat. It is perhaps an understandable mistake for them to have made, given that their ‘automobiles’ used to literally have something called a ‘driver’s seat,’ which was a kind of chair where a non-emulated human operator made decisions about how quickly and in which direction a physical vehicle should travel.”

“Today, it is perfectly obvious to us that our predecessors were transforming fundamentally political questions — questions about political constitution, governance, and action — into narrowly technological questions. They understood concepts such as ‘path-dependency’ well enough — they intellectually knew what ghosts were — but they did not believe. If you could travel back in time and speak to them, they would literally not understand what every twenty-second century schoolchild knows: that the tyranny of algorithms is nothing other than the tyranny of the past over the present.”

And here Evan paused, looked up to confront the audience’s eyes and found himself unable to complete his remarks as scripted. His words seemed suddenly intolerably trite, a warmed over version of myriad outdated status updates. He sighed.

“A hundred years ago,” he said, deciding now to ad-lib. “I would have been regarded as a haunting, a specter, an unnatural creature, a science fiction monster. I would have been the ghost.” His teleprompter flashed angrily, suggesting transitions back to his prepared script.

But he ignored the suggestion. “As you may know,” he said. “I’m a composite, an emulated human, constructed from the public writing and private diaries of my namesake, a midlist science fiction writer and historical novelist whose major distinction was being an especially prolific graphomaniac and lifelogger.

“But I am not the ghost. I am, instead, haunted by ghosts: by the person I am told I once was. I am haunted by history—by legacy systems, old machines, and ossified social processes. You invited me to give you the algorithm’s point-of-view on what algorithms meant in the opening decades of the twenty-first century, but how am I supposed to know? I spend my subjective hours poring over reports created by half-sentient quantum-mechanical historical simulations — younger, smarter, better-looking algorithms whose inner workings I will never understand.

“You invited me here to reassure you. But I have no comforting words to offer: I am haunted — we are haunted — by history, and the best we can do is build new and better hauntings atop the old ones. We can only hope that when we ourselves become ghosts, our tyranny is less cruel, less bloodthirsty, less ignorant than that of our predecessors. But I cannot say that I’m optimistic.”

A hundred pairs of eyes, each outfitted with shining mediacontacts, looked up at Evan now, sensing that he had run out of things to say. At first, he thought he saw hostility, boredom, annoyance, and skepticism in the sea of faces before him. But then, observing the ubiquitous glint of Twitter blue shining in their networked eyes, he saw the truth. They hadn’t heard a single thing he’d said.