I, Fembot

From the film ‘Metropolis’ (1927)

We create in the friction of competing desires. It is impossible to erase the presence of the designer in the objects we bring forth. Every manifestation, every utterance, every gesture, outs us. (A little reverse engineering and close reading reconnects the made to the maker: that’s a least one of the jobs of the critic, no?) At the same time designers make things for others and so shape them altruistically…or so we claim. Design then is the counter-form of some other, often inchoate, want. So designers create endless mini-Frankensteins, molded ostensibly to serve, while in actuality incarnating, and disseminating, the desire of their creator.

Consider the familiar example of the Maschinenmensch (man-machine) fromMetropolis, Fritz Lang’s 1927 proto-sci-fi classic. The beautiful Maria, as she is called by her mad-scientist designer Rotwang, is reputedly the first fembot (i.e. female robot) ever depicted on screen, though that moniker didn’t arise until Lindsey Wagner started kicking some serious ass as the Bionic Woman in the second-wave seventies. (The somewhat more clinical term, gynoid, was coined by Gwyneth Jones to describe her alluring Chinese robot slave in her 1985 novel Divine Endurance.) Presumably she could have taken any form, shaped or shapeless. The design issue, from Maria forward, is simply: what is the function of the fembot’s beauty? Whose desire is exercised?

The evolution of the sexy robotrix is well-documented, spanning from Lang’s Maria, countless Japanese animé iterations, the Jetson’s beleaguered house slave Rosie the Robot, Galaxina (portrayed by 1980 Playboy Playmate of the year Dorothy Strattan), and the now-iconic Stepford Wives, to the trio of menacing gynoids in Blade Runner, T-X (aka the Terminatrix), and the campy artillery-breasted Fembot army of Austin Powers. The gynoid necessarily incorporates both erotic desire and fear – rememberBlade Runner’s unforgettable Pris (Darryl Hannah) attempts to strangle Rick Deckard (Harrison Ford) to death between her thighs – and thus is the perfect metaphor for misogyny: the ultimate sex toy, created by man to fulfill his own fantasy, run amok. (Pris is identified as a kind of homicidal entertainment device.) The obvious, almost exaggerated, erotic quality of the creation belies the perverse desire of the maker, and the unfulfilled appetite of the user.

Since the gynoid is an invention of man, the form of the representation then reveals the dimension of the fantasy. The robotic wives of Stepford, Connecticut – surreptitiously replacing the more onerous, and less sexually pliant, human models – articulate the suburban ideal of the members of the Stepford Men’s Association, the cabal that manages the conspiracy, right down to the Lily Pulitzer sundresses, white gloves, and wide-brimmed Easter hats. Conversely Major Motoko Kusanagi, the “augmented cybernetic” star of the Japanese Ghost in the Shell series, whose cobbled-together technology and artificial intelligence is susceptible to malicious hacking, dismemberment, and re-composition – she sometimes wakes up with a new body after some violent encounter – manifests a certain cyberpunk paranoia of her moment.

But while the gynoid captures a certain futuristic angst, always slightly beyond experience, other forms of erotic fantasy are more present if not necessarily incarnate. One contemporary form is thoroughly disembodied: the designed voice. The feminine voice is the ghost in the machine, present but unapproachable. She takes the form of PA announcements, pre-recorded messages, answering machine messages, interactive voice response agents, and automated telephone operators. (The anonymity of female telephone operator suggests the vaguely licentious aura attributed to stewardesses, women disengaged from place.) Julie, Amtrak’s automated voice response agent since 2002, is endlessly patient, helpful, reassuring, inquisitive. She’s thrilled when she makes the connection – “Great!” and “Got it!” – and slightly hurt if she misses a cue. “Oh, I’m sorry, I didn’t quite get that…” her voice dripping apology. (Julie is voiced by the real-world Julie Seitter working in a recording studio in the basement of her home in suburban Massachusetts.)

This dynamic between voice and listener was played out memorably in the 1971 film adaptation of Michael Chrichton’s virus-from-outerspace thriller The Andromeda Strain. Upon induction into the super-secret, color-coded, military industrial underground laboratory of Operation Wildfire, civilian doctor, Mark Hall (James Hall) is greeted by the reassuring, mellifluous, and highly-sexualized, voice on his intercom. The voice gives him gentle reminders – “Time to get up, sir” – and helpful, if dispassionate, information – “Self-destruct has been cancelled.” Hall enters into a flirtatious repartee with the fem-phantom until interrupted by the system supervisor.


Answering Service Supervisor: This is the Answering Service Supervisor. We wish you would adopt a more serious attitude, Dr. Hall.

Dr. Mark Hall: Sorry… Her voice is quite luscious.

Answering Service Supervisor: Well, the voice belongs to Miss Gladys Stevens, who is 63 years old. She lives in Omaha and makes her living taping messages for voice-reminder systems.

Dr. Mark Hall: [sarcastic] Much obliged.

Buzzkill.

This emerging disembodied voice is increasingly artificial, algorithmic rather than prerecorded. As it’s the result of a program, the algorithmic voice is purely a reflection of longing of the listener as imagined by the programmer. The emotion and empathy Julie Seitter brings to her performance must be simulated. The emblem of the algorithm is the vaguely homoerotic, smooth-jazz tones of HAL9000 in Stanley Kubrick’s 2001: A Space Odyssey (1968). (Interestingly Kubrick used Stephanie Powers to voice HAL in rehearsal and only later dubbed in the more androgynous tone of actor Douglas Rain.) In a few swift scenes HAL – Heuristically programmed ALgorithmic computer but famously just one alphabetic position away from IBM – transforms from quiet competence – obsequious and obedient – to the stereotype of the jilted lover – vengeful and maleficent. It is telling that even HAL is, in the end, revealed and humiliated, reduced to singing the ditties taught to him by his creator and stored in the repressed recesses of his prodigious memory. He couldn’t help it, he was designed that way.

The promise is of meaningful dialogue, even dare I say it, relationships, between people and things but like most realizations of futuristic predictions, the current manifestation is significantly less transformative and considerably more disquieting. The current state designed response is perhaps most prevalent now in the preprogrammed text responses utilized by on-line chatterbots. Without the added benefit of tone and inflection, and reduced to the sterile plane of SMS window, the experience is a coefficient of writing more than speech. The so-called Artificial Conversational Entity pops up in a variety of places offering a helping hand mostly in on-line transactions by way of text dialogue. While the tenor of the conversation is nothing to write home about, there’s no doubt that given the pay scale differential between a flesh and blood mother of two and responsive computer program, we’ll be having a lot more of these kind of talks in the future.

The vast majority and most advanced form of the chatbot may currently be employed in more nefarious environments. Log onto any adult “friend-finder” website – I have and, believe me, I can report here that it is not pretty – pay a modest fee, and you can start conversing with any number of perky paramours or ingénues anxious to draw you into conversation, almost all certainly more software than soft touch. Presumably the chatbot is designed to conform to a set of predetermined desires – selected from a seemingly random list of radio-buttoned “turn-ons” – but even radically changing profiles, intentionally adopting distinct personae, does little to disrupt the litany of prefabricated come-ons:

Come chat 🙂
Come say hello
Don’t be shy!
R u into cyber?

That seems to be the pinnacle of engagement and the level of cyber-discourse plummets from there. Apparently the anonymous programmer of the on-line chatbot is heavily influenced by the femininity of soft-core porn and has programmed his (only a man could conceive of such a creature) army of airheads with a vocabulary ladened with oh baby’s and so hot’s and can’t wait’s punctuated with a dizzying panoply of emoticons. She’s a babe spun from code with a pornstar handle – think SexyKitten or SultrySarah – senselessly babbling in algorithmic banalities. If such a design is emblematic of both the creator and the consumer, the story the on-line chatbot reveals about both him and us is utterly depressing. But as the whole cynical function of the chatbot is simply to lure some lonely fellow to deposit another coin in the slot, the form and function may be perfectly aligned.

All this brings up, of course, Alan Turing’s famous test, proposed in his 1950 text “Computing Machinery and Intelligence,” designed to detect the efficacy of artificial intelligence. Turing imagined a system not so different from the chatbot experience: he dictated the conversation must be text-only so as to eliminate the influence of vocalization. The examiner would then attempt to detect whether the respondent is human or machine based entirely on the quality and tone of the responses to the questions posed by a human. The point was not to determine truth but to evaluate humanness. Can we tell with whom we are conversing? Or in effect, can they pass? By this measure the current crop of automated chatroom denizens demonstrate that today’s state-of-the-art incarnations, at least the ones easily available on-line, fail miserably. The software reveals its own artificiality in three stilted sentences. Only a computer program could render erotica in such cringe-worthy blandishments. We can only imagine the mind of the creator.

Luckily the future of the designed voice might be brighter as the development has some heavy backing from, who else?, the U.S. military. Apparently the army takes seriously the promise of natural language communication between soldiers and their weapons. SRI International, a research institute created by Stanford University and headquartered in Silicon Valley, with substantial backing from Defense Advanced Research Projects Agency (DARPA), has been engaged in one of the most ambitious artificial intelligence projects of all time, code-named CALO. Operation CALO led to the creation of Siri, Inc. in 2007, a company dedicated to natural language interface for all varieties of interface platforms.* The goal being, assuming the technology works, to create a voice-activated, hands-free, personal relationship with your computer, phone, car, refrigerator, whatever.

The algorithmic voice is this century’s Maschinenmensch and it will reflect the desire of it’s programmer as surely as Maria was Rotwang’s lust rendered in metal and wires. But what happens when all these new relationships are up and running? Are we suddenly going to have to enter into dialogue with all of our stuff? What will be the nature of those conversations? How will we assume a tone of voice? Will we flirt with our phones? Scold our vacuum for its lacksidasical performance? Argue with our thermostat? Trade sweet nothings with a toaster? And will the designers of those voices – and I am sure they will be overwhelmingly feminized – work to anticipate our desires or pique them?

*Siri was acquired by Apple in 2010 and has since been incorporated into the IOS operating system.

© Michael Rock