Cyborg Intelligence: Dismantling Gender in the Age of Chatbots

Words by A.E. Zimmer

Bionic Hand Prosthesis

A few weeks ago you would’ve thought some transhumanist war had been launched with two words: Eugene Goostman.

Eugene Goostman: smart name for the smart robot crowned champion of the infamous Turing Test, the standard for determining machine intelligence created by British mathematician and computer scientist Alan Turing.

At the 2014 Turing Test At The Royal Society in London, the robot, under the guise of a 13-year-old Ukrainian boy, answered questions about his hometown of Odessa, his Jewishness, his pet guinea pig, and his father, a gynecologist (“My friends say he is a “beaver-doctor” – but I guess they lie – he is not a veterinary!”).

After successfully convincing 33% of visitors that he was indeed human, Eugene was hailed as a long-coveted, shining bit of proof: at long last, our technologies were approaching fantastic, sci-fi proportions.

Media outlets lit digital torches, sounding the news of this cyborg heir apparent, only to find out later that by most qualifications, this star pupil Eugene had actually failed.

The reason for our star pupil’s flunking? It turns out Eugene’s satisfying performance was less about his abilities to “think” intelligently , than it was about his programmer’s abilities to account for his flaws. The bot’s programmers admitted that the description of Eugene as a 13-year-old with “a boyish personality”, with English as a second language, greatly helped convince judges of the robot’s humanity. Such a description allowed for more human “forgiveness” for Eugene’s mistakes of syntax, grammatical errors, misspellings, and the like.

Blunders aside, what’s perhaps most interesting about Eugene’s personality is the decision to call him male. After all, what does it mean when the most “intelligent” machine “passes” as a boy? How important is it that when a machine pretends to be a young, (presumably straight) white male , it is lauded for its success and intelligence?

Eugene Goostman is a chatterbot– the kind of algorithm that made you waste hours on AIM with Smarter Child, or Cleverbot. They are systems of relatively simple strategy: programmers wire robots with lists of potential phrases and through matching these phrases to data, generate a response. They also frequently use techniques like phrase repetition as well. Algorithmically speaking, there have been plenty of chatterbots equally capable, as Eugene in “faking human” that’ve received far less hype and adulation. This is particularly true of those bots gendered female.

Consider ELIZA (note the Pygmalion namesake), a famous 1964 program modeled after your most infuriating psychotherapist. You can even ask SIRI about her.

Though her programming logic was fairly simple, ELIZA was revolutionary for her appeal with humans , as many of her “patients” reported growing attachment to the bot. While its newfound popularity pinned ELIZA as a contender for passing Turing’s Test, her victory never came to fruition. Eventually her programming was used in new chatterbots with new names, some of them gender neutral, most of them male-identified. Before you knew it, that was the last “female” chatbot given any shred of Turing Test credibility.

It only takes a quick Google Search to see how the second-wave feminist chatbots have fared: There’s Athena, the loftily-titled GoddessBot, a “nice and unpretentious” french bot named Jane, Mitsuku, “your new virtual friend… here 24 hours a day just to talk to you”. You catch the idea.

None of these chatterbot programs convince humans of much of anything other than some validation for their latest fetish, yet Eugene’s “personality” absolved him of these shortcomings. It’s not that the Ukrainian bot was smart enough to pass the Turing Test, but–unlike its feminized counterparts– the qualities of its imagined gender, age and sexual identity allowed the robot’s “passing” to the public.

What’s the value of gendering our Artificial Intelligence? Perhaps it’s a programmer’s matter of habit, a partiality to A.I.’s history of a “neutral” standard that’s categorically male. And yet, A.I.’s history has never been “neutral”– it is implicitly gendered, and implicitly queer–Alan Turing himself, godfather of Artificial Intelligence, led his life as an openly gay man of the 1940s. The very concession of his queerness found him criminally prosecuted for homosexuality, subjected to horrific violence that would lead to his suicide: one apple bite, laced with cyanide.

Many A.I. enthusiasts rarely, if ever, entertain the truth of their hero’s life. Similarly, the gender politics at the core of Turing’s original “test” are often disregarded–which of course is no test at all, but something entirely different.

Jolie

“Turing never proposed a test in which a computer pretends to be human,” says Karl MacDorman, associate professor of human-computer interaction at Indiana University. “Turing proposed an imitation game in which a man and a computer compete in pretending to be a woman. In this competition, the computer was pretending to be a 13-year-old boy, not a woman, and it was in a competition against itself, not a man.”

Though it’s no secret that tech has a gender problem, what’s makes A.I.’s problem with gender particularly hard to stomach is how this study tends to thrive on imagination. Robots, after all, are ridiculous things, stuff of dreams and nightmares. In them we plant very human hopes and fears for the future– the considerations of robot bodies and the constructions of robo-minds are often chrome negotiations of our human lives and their fleshy experiences.

So when the “smartest” machine in the world is a 13-year old boy, cracking vulgar genital jokes about his gyno dad, what fiction are we selling ourselves? When the very inventor of machine consciousness was an openly gay man, chemically castrated for this very admission?

Donna Haraway’s Cyborg Manifesto– seminal resource for any analog girl really out here in a digital world– confirms the cyborg as the image of so many competing concepts:

“The cyborg is a condensed image of both imagination and material reality, the two joined centres structuring any possibility of historical transformation. In the traditions of ‘Western’ science and politics — the tradition of racist, male-dominant capitalism; the tradition of progress; the tradition of the appropriation of nature as resource for the productions of culture; the tradition of reproduction of the self from the reflections of the other — the relation between organism and machine has been a border war. The stakes in the border war have been the territories of production, reproduction, and imagination.”

As this “border war” plays out, what productions and imaginings are denied, limited, and silenced in A.I. research? When the image of intelligence is an android faking boy, how does this reflect on our definitions of human intelligence?

While A.I. offers an outlet for unbounded fantasies, so too is it an academic field born out of an obsession with dualism. The robot, at its most glorious, is one big, glistening oppositional binary, the Cartesian affirmation of the mind-body catch-all “I think therefore I am.” For anyone worth their weight in sexual politics, it’s well understood that a commitment to dualistic thought–mind v. body, him v. her, us. v. them– is the first step toward brutalities of all kinds.

After all the cloudy discourse, Eugene Goostman’s hype paints a message that’s crystal clear: when we connect our ideas of intellect to a machine called “male”, we’re meant to connect the idea of intelligence as an inheritance of men.

As long as we actively align ideas of consciousness with male-identified presentations of consciousness, we invest in a future where sexual politics resemble that of a Transformers sequel. Without the authority of a male gendered voice, all we’ve invented is a plaything, a “friendly” chatbot, a fembot, a Siri, a personal assistant that can lead us to the nearest escort service in the area but never the nearest abortion clinic. This is of course, to say nothing of the racial and sociopolitical binaries of tech at large– hopping down that route would require a whole ‘nother rabbit hole entirely (with a 5,000 word minimum).

Though singularity might be some time away, it’s prudent to remember these are active, present, virile narratives. When whole universes of experience are left out of the story of the future– however inventive, speculative, or silly– we take steps toward deleting the identities and bodies that live them.

Fortunately, the future is ripe for dismantling glitches of history. We would do well to dream big in these days of transformation, all while keeping one watchful, too-human eye on bots, and the bros who build them.