Do we need to revert to traditional constructed gender binaries in order to relate to artificial intelligence?
Is there any reason to expect that technological advances will address gender inequality?
Donna Haraway raised this question in her 1991 essay, ‘A Cyborg Manifesto’. Here she pointed out that ‘the main trouble with cyborgs, of course, is that they are the illegitimate offspring of militarism and patriarchal capitalism.’ She also noted, however, that ‘illegitimate offspring are often exceedingly unfaithful to their origins’, and while they can be manipulated and anthropomorphised towards our own designs, the image of the cyborg is also a rejection of rigid boundaries.
Robot technology has always forged ahead according to human expectations. When smart artificial intelligence has appeared far-off, it has failed to convince backers, causing ‘AI winters’ (a period of time where there is reduced funding and interest in artificial intelligence research). But when it has seemed susceptible to misbehaviour by acting of its own accord, it allures and is funded. When artificial intelligence tricks human beings into thinking they are interacting with their own – which is often the endgame – the scales are tipped. Are humans machines, and machines humans?
One of the earliest feminine-gendered AI (also known as a ‘fembot’ or a ‘gynoid’), Actroid, was built in 2003. First programmed to smile and flutter her eyelids, models were subsequently developed to function as receptionists, waitresses and cleaners. The latest version, Actroid-F, fooled test participants into thinking she was a real woman, when placed side-by-side with one. Commenters watching Actroid-F on YouTube described her as ‘sexy’.
Previous, rudimentary prototypes have preceded the likes of Actroid. In 1983, the robot-cum-coffee machine, Sweetheart, ran into controversy when Berkeley feminists petitioned for it to be removed from an exhibition.
A five-foot-tall robot, Sweetheart was made up of recycled metallic objects with baseball-bat legs, a coffee pot head, a coffee urn torso, and two desk-lamp shades as breasts to illuminate a busty silhouette. When interviewed, Sweetheart’s creator Clayton Bailey said it was ‘the world’s most beautiful lady robot,’ and to ban it was akin to condemning the female form itself.
The idea that the ‘perfect’ woman is an artificial one isn’t anything new; in fact, it’s a common trope. From Amazon’s digital assistant Alexa, to female-appearing bots built as domestic servants and sexual slaves in Westworld, to Ricky Ma’s Scarlett Johannson-lookalike gynoid Mark-1, and to Scarlett Johannson herself as Samantha in Her, these AI have shown themselves to be constantly available and always supportive.
Like the hospitality worker in most service roles, AI are taught to exercise emotional labour where they are required to display poise and deference to their human equivalent. The connection between the high percentage of women as carers, sex workers, secretaries and housekeepers, and gynoids built to perform these roles isn’t a mere coincidence; the popularity of feminine-gendered AI is evidence of a world where women still aren’t seen as fully human.
Like the tale of the sculptor Pygmalion, the relationship between humans and artificial counterparts dates back to the myths of ancient Greece: Pygmalion had fallen in love with his statue, bringing it to life with a kiss after wishing for a bride who would be the ‘living likeness’ of his ‘ivory girl’. Interestingly enough, the first chat-robot, Eliza, was named after Eliza Doolittle in George Bernard Shaw’s play, Pygmalion. What better parallel than the narrative of a working-class Cockney girl transformed (re-programmed) to fit in with elite society?
Like Pygmalion, the relationship between humans and artificial counterparts dates back to the myths of ancient Greece.
History continues to repeat itself even as technology makes bigger headways in this brave new world. In what Haraway calls the ‘informatics of domination’, artificial intelligence continue to be coded as female, in a world where men innovate and build, and women backup and deliver (only 26 per cent of women hold computing jobs, and five per cent of leadership positions in the tech industry are held by women).
From what we have seen with current working AI such as the likes of Apple’s Siri, Microsoft’s Cortana, Amazon’s Alexa, and Google’s Google Now, these prominent bots are designed to sound like women while they helpfully obey our commands. These ethical quandaries often reflect easy solutions decided by the male-dominated tech world. As AI user experience designer and writer Jacqueline Feldman notes in the New Yorker, ‘the method for insuring that a technology speaks without giving offence has been to make it a woman, even if some designers later change their minds.’
What does this mean for gender equality in a post-tech future? When technology evolves to mimic the advanced prophecies speculated by an ostensibly less progressive past, while the status quo dictated by societal biases continue unshifting, the gap feels like a discrepancy. Feminine voices will remain more commonplace because people tend to perceive them as assistants that help us to solve our problems by ourselves, whereas masculine voices are viewed as authority figures that tell us the answers to our problems.
In a study conducted by a group of researchers at the Indiana School of Informatics, it was found that when it came to automated voices, people generally preferred female voices than male ones. In another study, male participants were significantly more likely to rule against computer-generated female characters when her movements were jerky, while this had no impact on female participants.
People tend to perceive [feminine voices] as assistants that help us to solve our problems by ourselves.
Visions of the future persist in propping up this kind of gendered stasis even in our cultural canons. Familiar cinematic and literary trajectories often follow the figure of the once-compliant fembot gone rogue, while the audience grapples with the conundrum as to whether she should be reined back in and stopped.
It’s the damsel-in-distress trope with a twist: instead of a hero coming to save her, the sexy fembot saves herself from herself, but ends up destroying everything around her in the process. A hat-tip to the caricature of the female psyche, the message is clear: rebel and risk loss. Sometimes, like Pris Stratton (‘a basic pleasure model’) in Blade Runner and Buffybot in Buffy The Vampire Slayer, the bot isn’t given an option at all.
The gendered implications of artificial intelligence are often so buried in the gender binary present in our human societies that it’s currently still difficult to isolate the two. To the tech industry’s credit, bots constitute a big leap in technological development. In an artificially intelligent world, debates surrounding the ethics behind the domestic and emotional labour once relegated to humans we view as less privileged in the class system are gradually falling away.
But if bots continue to echo the remnants of a sordid past by bringing its attendant politics into the present, and eventually the future, the lines between ‘male’ and ‘female’ don’t have an opportunity to blur but instead become starker and starker. If we don’t see our tech as clean slates wherein we are given the opportunity to radically alter deeply ingrained societal patterns, where else are we likely to falter?
We don’t need to make our tech acquiesce to a forcibly constructed gender binary in order to relate to them. The anthropomorphic impulse seems like an obvious conclusion in a world where we view animals in our likeness and attach emotions to lifeless objects. But when AI are attributed personalities in a self-serving quest for human likeness new guidelines are needed. Can a robot be personable, helpful and clever without the accompanying baggage brought over from a flawed society?
As Donna Haraway puts it, ‘This is the self feminists must code.’ The first feminist robot may not necessarily be a woman. The first feminist robot will be able to consent.