Sam Visits the Uncanny Valley

Sam, in younger, more innocent days

It’s hard to escape the uncanny valley these days, even if you’re a dog. Our puppy Sam learned that lesson during a visit to the vet last week.

A robotics professor coined the term “uncanny valley” to describe the discomfort we feel when we perceive that something inanimate is trying too hard to convince us it’s alive. As Wikipedia puts it, “when human features look and move almost, but not exactly, like natural human beings, it causes a response of revulsion among human observers.”

Sam’s experience at the vet made me think that definition may lean a bit too heavily on the “human.”

Sam is a six-month-old yellow lab, and she’s a sweetheart. I can literally count on one hand the occasions when she’s barked or growled at anyone or anything.  None of those occasions happened at the vet, which is why we were so surprised last week when Sam let out a couple of barks and started growling as soon as we took our seats in the waiting room. 

What set her off, it turned out, was a decorative wooden dog in a corner of the waiting room. It was a blocky, slightly comic version of a terrier, about two feet high, with a combative expression on its face and a colorful scarf around its neck. Sam was deeply suspicious. The fur on her back raised and her lip curled. Instinctively she seemed to know that something wasn’t right here – that this “creature” was trying to pull a fast one. Sam had entered the uncanny valley.

I know how she felt. We humans are having our own trouble figuring out what’s real and what isn’t these days. Those doubts are, in turn, symptoms of more general anxieties about where we stand in relation to machines.

We live in an age when technology is ascendant, and we’re thrilled by all the things our devices are doing for us. Our enthusiasms, though, are accompanied by an undercurrent of fear. Technology gives us power, but we know that power cuts both ways. Subconsciously we’re aware it can turn against us.

This is the essence of a theory called the Fourth Discontinuity, which derives from a comment of Sigmund Freud’s. Through most of our history, he said, human beings were confident the universe revolved around them. In the past 450 years, however, that cherished self-image has suffered three major blows. The first of these was the Copernican Revolution, when we learned that Earth is a satellite of the Sun, rather than the other way around. The second was Darwin's On the Origin of Species, which showed that man was descended from apes. The third, Freud modestly contended, was his theory of psychoanalysis, which demonstrated that our thoughts and our behaviors are not entirely in our control, but instead are influenced by drives and conflicts hidden in the deepest regions of our personalities.

 

The fourth discontinuity is a recent addition to the list that takes into account how quickly machine intelligence has advanced in the past fifty years. Just as the second discontinuity acknowledged that we can no longer claim to be an order of nature distinct from and superior to animals, the fourth discontinuity holds that we can no longer claim to be an order of nature that is distinct from and superior to machines. The result is an ongoing identity crisis.

Our feelings toward animals today have a lot to do with our misgivings about technology. In his book, The Metaphysics of Apes: Negotiating the Animal-Human Boundary, Raymond Corbey notes that throughout history there have been periods when cultures have emphasized the differences between humans and animals and periods when they’ve emphasized their similarities. The publication of Darwin’s theory provoked a swing of the pendulum toward discontinuity; people wanted to prove how absurd it was to suggest our ancestors lived in trees. Today the pendulum has swung decidedly in the opposite direction. Not a day goes by, it seems, when there isn’t a new book, article or study talking about how much humans share in common with elephants, dolphins, chimpanzees and dogs. Especially dogs.

Animals serve as antidotes to the Fourth Discontinuity. If our interactions with technology make us feel, on some level, inferior, it makes sense that we take comfort in our kinship with animals. Sam’s behavior at the vet can be counted as another sign of our affinity. Together we stand on one side of the uncanny valley; robots and wooden terriers are on the other. Dogs are indeed man’s best friend, now more than ever.

###

(An earlier post on creepy technologies addresses related issues.)

©Doug Hill, 2013

Creepy Technology

The online magazine Slate ran an essay this week that asked the question, "Why Do We Love To Call New Technologies 'Creepy'?" The article was written by Evan Selinger, an associate professor of philosophy at Rochester Institute of Technology.

My initial reaction to that essay, posted here, was critical, but Selinger suggested in a tweet that I'd missed his point, which he said was "'creepy' discourse + normative analysis."

He's right, I'm sure, that I missed his point – to be honest I don't know what "normative analysis" is. So, with apologies to Selinger, I've reworked the essay to ask, simply, What is it about some technologies that makes us feel creepy?

There's an obvious correlation between creepiness and novelty. It's not unusual to be suspicious of strangers, especially when they have the potential to effect some degree of change in our habitual sense of the world. With technologies as with people, a measure of trust has to evolve.

Selinger's essay mentions that early railroad passengers sometimes developed a variety of symptoms that physicians came to recognize as manifestations of "train sickness." He suggests these maladies were a reaction to the creepiness of unfamiliarity, a form of "mania" that simply disappeared with time. Without going into detail (or normative analysis), it's worth noting that the experience of early train travel was considerably rougher and more dangerous than it would become as technologies of comfort and safety evolved. 

Still, I don't doubt that (to borrow Robert Hughes' phrase) the "shock of the new" had something to do with passengers' uneasiness. As a given technology weaves its way into our lives the creepiness factor usually fades, as does its "specialness" factor. We become acclimated to its presence, and then dependent on it. The miraculous and frightening become routine. I say the creepiness factor "usually" fades because it doesn't always. Plenty of people still find flying on airplanes creepy, for example. I'm one of them. Still, during periods in my life when work required frequent air travel, creepiness faded and acclimation set in.

There are two less obvious issues that help explain the creepiness we often feel in response to technology. One is that we're intuitively aware that, in terms of brute strength, technological power outstrips human power. You don't have to be a religious fundamentalist to see that technology is about achieving a degree of mastery over nature, other human beings, and ultimately death that was once believed to be the exclusive purview of God. But we're also aware that technological power cuts both ways, and thus is not only a source of security, but also fear.

Human vs. Machine (and Machine/Human) in Desk Set

Immediately after reading Selinger's essay I happened to catch on TV a showing of the old Katherine Hepburn/Spencer Tracy movie, Desk Set. The plot, for those who don't know it, revolves around a group of women who staff the research department of a major television network. Much of their working day is spent answering questions from the public – someone wants to know who had the highest career batting average in the history of baseball, someone else asks for the names of the reindeer in "The Night Before Christmas." The women handle these calls with an impressive combination of dedication, good humor, and smarts. 

The conflict in the picture is supplied by the installation in their department of a new, room-sized computer named EMERAC (a variation of the names of the early computers UNIVAC and ENIAC), which the researchers assume has been brought in to replace them. It turns out (spoiler alert here) that the movie isn't really about the threat of automation – that's just an excuse for a romantic comedy that revolves around Hepburn's researcher falling in love with Tracy's efficiency expert/computer engineer. It's a formula that requires a happy ending, and indeed, in the end we learn that the researchers aren't fired, EMERAC is only there to help them.

That the plot goes in this direction perhaps explains why there's a note in the film's opening credits thanking IBM for its assistance in the production. In any event, by the end, Hepburn's character (who's named, surely not by coincidence, Miss Watson) is cheerfully learning to use the computer and cooing affectionately as it spits out answers to questions. Human and machine learn to live in mutually supportive collaboration; mistrust and fear give way to admiration and gratitude; the researcher accepts the engineer's proposal of marriage.

This seems a bit disingenuous. In truth, there's good reason for the researchers of Desk Set to fear the arrival of EMERAC. Countless workers, from the onset of the Industrial Revolution to today, have been replaced by machines. Technology has the power to make us obsolete, and we know it. That's creepy.

A merger of man and machine goes poorly in David Cronenberg's The Fly

The second fundamental issue that the creepiness question raises is an existential one. It involves the alienation that exists between two separate orders of being: the organic and the mechanical. That's what the uncanny valley is about, I think. We instinctively recognize that a machine is trying to sneak across that boundary, and it puts us on our guard. A similar discomfort may be at the root of the creepiness some people feel about flying: there's just something unnatural about it. 

Desk Set gets lots of mileage out of this tension. The computer and the female technician who's brought in to attend it are both portrayed as cold, relentless intruders into a human community. Stanley Kubrick played brilliantly on that tension, too, defining the character of HAL in 2001 with two radically incongruous features: a heartless, staring eye and a voice that oozed creamy sincerity.

I realize not everyone agrees that a firm line exists between human beings and technology. There's no reason, many believe, we can't share ontological space with one another. Certainly the transhumanist point of view is that nothing could be more natural than humans merging with their machines. “[I]t is our special character, as human beings, to be forever driven to create, co-opt, annex, and exploit nonbiological props and scaffoldings," writes Andy Clark, author of Natural Born Cyborgs. "…Tools-R-Us, and always have been.”

I don't buy it, or, more accurately, I don't buy the implication that such adaptations are necessarily desirable. Human/machine intimacy is as likely to produce mutation as it is enhancement, in my opinion. This is a view that draws me to the work of artists like David Cronenberg and Philip K. Dick and philosophers like Jacques Ellul and Herbert Marcuse.

One of the more eloquent expositions of this perspective came from the theologian Paul Tillich. Like other existentialists, Tillich believed that uneasiness is endemic to the human condition. It's weird being aware that we exist and weird knowing that we're going to die. Our predicament leaves us with persistent feelings of, as Tillich put it, "uncanniness."

We've come up with lots of ways to avoid those feelings, and technology is high on the list. On one level we find technology reassuring because we think we can control it.  Even though we may not understand how it works, we believe it behaves, Tillich says, by rational, logical, "calculable" rules. We can surround ourselves with it, cloak ourselves in it, and feel secure. Tillich cites the home as an example. Its “coziness,” he wrote, holds “the uncanniness of infinite space” at bay. What the house or apartment offers individuals, the city offers humans en mass.

Like so many palliatives, however, technology can turn on us. It may not be as safely in control as we'd hoped. The potential for unease grows as our technologies become more powerful, more complex, and more self-determined. On some level we're aware that the relentless logic they're following is their own. We know they're not truly alive, but they seem to be. We wonder whose agenda is being followed. Creepiness ensues.

"As the technical structures develop an independent existence," Tillich wrote, "a new element of uncanniness emerges in the midst of what is most well known. And this uncanny shadow of technology will grow to the same extent that the whole earth becomes the 'technical city' and the 'technical house.'"

Tillich ends this passage with a pertinent, and creepy, question: "Who can still control it?"

©Doug Hill, 2012