Alan Turing is one of those geniuses whose moment emerged decades after his death. Yet even as our estimation of Alan Turing’s contributions and brilliance continues to grow, the legacy of his eponymous test is more complex. A staple among journalists pitching stories about bots and AI’s encroaching on the human, the test itself doesn’t hold up very well to scrutiny. It’s never really been clear exactly how the test is supposed to work. Taking Turing’s thought exercise into the popular consciousness and making it the goal of artificial intelligence has involved several elisions and simplifications of the original. It’s become the stuff of good theater and journalism, but not a great metric for the progress of technology.
I recall coming to this realization in 2011, when I read Brian Christian’s long piece “Mind vs. Machine” in The Atlantic outlining how much strategizing goes into being human, i.e., into figuring out how to act in such a way as to differentiate yourself from a machine. Indeed, Christian tells the story of preparing himself like an athlete, analyzing game tape of previous tests, and managing to distinguish himself not just from a machine but from the other humans as well, ultimately winning the “Most Human Human” award.
Our doubts about the utility of his test, though, have been accompanied by growing appreciation for deeper meaning of his fascination that inspired it. Turing’s test was a riff on the Imitation Game, a parlor game originating in a Victorian curiosity about whether a recipient could distinguish a man and a woman writing as their biological sex from a man or a woman attempting to pass as the opposite sex.
As Clive Thompson writes in Wired (“The Original Turing Test”)
“As a gay man who spent nearly his entire life in the closet, Turing must have been keenly aware of the social difficult of constantly faking your real identity. And there’s a delicious irony in the fact that decades of AI scientists have chosen to ignore Turing’s gender-twisting test.”
Add to this the values of the day, as well as the speculation that Turing was somewhere on the autism spectrum, and one imagines that he might have seen black and white lines in places where we’ve come to see more fluidity.
It’s no surprise that a literary and cultural theorist of the cybernetic like Katherine Hayles would also find meaning in the fact that it’s not just a contest of human versus machine, but also of gendered bodies versus each other versus machine.
Think of the Turing test as a magic trick. Like all good magic tricks, the test relies on getting you to accept at an early stage assumptions that will determine how you interpret what you see later. The important intervention comes not when you try to determine which is the man, the woman, or the machine. Rather, the important intervention comes much earlier, when the test puts you into a cybernetic circuit that splices your will, desire, and perception into a distributed cognitive system in which represented bodies are joined with enacted bodies through mutating and flexible machine interfaces. As you gaze at the flickering signifiers scrolling down the computer screens, no matter what identifications you assign to the embodied entities that you cannot see, you have already become posthuman. (How We Became Posthuman xiv)
In Hayles’ posthuman, the cis- and trans- distinction breaks down in the exercise of attempting to detect that distinction, whether in the category of gender or material substrate. As in Christian’s behind-the-scenes account, you can’t help but know that everyone is trained (or training) to sound like someone, whether it’s their other or the essence of what they supposedly are. Not only are both acts of charade, they’re also programmed roles, in which we run routines and stratagems to communicate meta-messages about who we are and who we aren’t.
This last distinction leads to the deeper reason that we can’t quite figure out how a test like Turing’s is really supposed to work: who we are and who we aren’t has never been a simple matter. This isn’t a new problem—Aristotle, Augustine, Dante and Shakespeare (among others) had a lot to say about it—but no one quite turned everything on its head like Freud. For Freud, it’s not simply the fluidity of the self-other distinction that’s outed in this masquerade of self-hood; it’s also the divisions within ourselves—between conscious and unconscious; between the identities we embrace and the identities we deny. It’s no wonder that Freud is invoked in that cousin of the Turing test, the uncanny valley.
Commonly used to account for our fear and revulsion at the almost-human—a wax museum figure, prosthetic hand, or robot made to resemble a human—the uncanny valley simultaneously evokes and neuters the psychoanalytic origins of the term. It’s misunderstood to describe a simple human response to an external object. For Freud, the object evokes the sensation of the uncanny for a deeper reason. It’s easy to attribute the feeling to the object’s obscured-yet-detectible otherness—it’s not-quite-human, it’s not-quite-normal quality. But for Freud, the feeling can only aroused because it serves as a reminder of the internal otherness within us, the one that we’d prefer to obscure from ourselves. When we respond with fear and revulsion to stories of haunting and possession, we do so because we find ourselves haunted (and occasionally possessed) by the repressed aspects of ourselves—our uncivilized urges, our untamed desires, the parts of ourselves that we can’t countenance.
The literary and psychoanalytic consideration of the uncanny all starts with an analysis of E.T.A. Hoffmann’s short story, “The Sandman” , the story of Olympia, an automaton, a bot falling short in life’s imitation game. Her failure to pass as human (or to pass the Turing test), as well as our our aversive experience of that failure, prefigures many a science fiction plot, from Bladerunner to Her. For Freud, the almost human is spooky not because it isn’t human but precisely because it reminds us that we’re human precisely by virtue of our primitive, irrational and raw parts–parts that something with gears or circuitry at its heart can never fully embody. We might like to think these desires and urges belong entirely to the other—the monstrous, the criminal, the perverted—but they haunt us, and imbue monsters—zombies, werewolves, vampires—and the grotesque with a magnetism we can’t look away from.
Ultimately, when we’re trying to distinguish between man and woman, between the authentically human and the programmed-to-be-human, we’re trying to make a statement about who we are and who we aren’t. Even the cis- / trans- binary fails when our identities pass in and out of ourselves and others, both human and inhuman. Never has Rimbaud’s “Je est un autre” been more true: we project personality traits on our brands and our machines and we receive validation for our identity from devices in our pockets and media designed to hijack our attention. In this context, the Turing test is a cyborg’s anachronistic dream of being an authentic anything. But as much as its all-too-human to fantasize about a test that can distinguish real from fake (see Blade Runner’s Voight-Kampff test, it’s ultimately much more liberating not to be untethered from having to pass any test at all.
2 thoughts on “The Turing Test, Part I: The Imitation Game, Masquerade, and the Uncanny”