"With artificial intelligence we are summoning the demon. In all those stories where there’s the guy with the pentagram and the holy water, it’s like yeah he’s sure he can control the demon.
Didn’t work out.“1
There is a consistent trope exploring the artificial human:
Like Ovid’s Pygmalion Myth, in which Aphrodite breathes life into a sculptor’s creation, gifting him his object of desire.2Or with the Golem, a humanlike creature in Jewish literature and mysticism, mute and always obedient to its owner.3Or like the ‘false Maria’, the evil Doppelganger in Fritz Lang’s Metropolis, a femmebot fatale tricking everyone into believing she is her original.4Current depictions craft technological beings like the (almost) human hosts in HBO’s Westworld or Alex Garland’s Ex Machina’s emotionally manipulative Ava – the robots in them not only questioning the Conditio Humana but also their function.5All those human-machines have in common that they are created artificially, encased in a body of flesh, blood, silicone and/or metal. Ultimately, they seem to obtain consciousness (or are at least within its very reach).
The idea of creating life artificially – to break down the human mind – is indeed a long-lasting myth. But with technology, we seem closer than ever to its realization. Humanized robots are no longer the stuff movies are made of, they exist in real life, powered by Artificial Intelligence (AI) based systems. And AI plays a vital role in our life. Whether we are aware of it or not. The machine powered intelligences operate as cultural agents of taste, knowledge and ideology. As Lev Manovich states in his book AI Aesthetics, they are “(…) influencing our choices, behaviors, and imaginations.”6
Moreover they are recommending us who to know, to follow and what type of information we are getting when and how. As invisible, bodiless operators they lead the way we move in our world. But what happens if this artificially created ghost materializes in a human-made shell? Which demon are we summoning with that?
As Elon Musk states above, we are summoning the demon with AI. Despite Musk being a dubious agent of imperialistic ideology, I think his prophecy, although arguably exaggerated, is of importance. As a matter of fact, AI techno-capitalism has a lack of regulation, concluding in a black box of biased decisions (for which we can thank the ones making them), not able to be recalled by humans anymore.7Besides that, his statement also establishes a distinct holistic metaphor of AI as the ultimate demon bringing a technological Armageddon upon us. As we tend to picturize metaphors, one may ask: How does this demon look like? What if said demon looks like, let’s say maybe a tall, beautiful machine-woman? Ai-Da, as quoted by her creators, the gallery director Aiden Meller and curator Lucy Seal, is the world’s first ultra-realistic, humanoid AI robot artist. Named after the first computer programmer Ada Lovelace, she is an AI-powered robot, epitomized in the picture of a multidimensional artist: “She can draw, and is a performance artist. As a machine, with AI capabilities, her artist persona IS the artwork, along with her drawings, performance art and collaborative paintings and sculptures.”8She also looks like a very beautiful, light-skinned woman with long brown hair, full lips and empty eyes.
Responsible for her design was the robotics company behind the critically acclaimed HBO show Westworld. They used markers of humanness in giving her a face, hair and a neck. Unlike her upper part, the rest of her body is not covered in silicone and actively reveals her artificiality. In fact, Meller and Seal emphasize that she is not real.9This decision could be interrelated to the circumstance that robots who are the spitting image of human beings are mostly perceived as threatening and aggravating. This phenomenon is known as the ‘uncanny valley’, a concept by Masahiro Mori, describing that we tend to feel more positive and affine towards humanlike robots as long they don’t feel too human. That’s when they sink into the ‘uncanny valley’, the dead space between ‘barely’ and ‘fully human’. The response turns into refusal as they mimic life in an inanimate way like zombies or corpses. Only when they are no longer indistinguishable from humans, they get accepted again and humans can feel empathy for them.10Ai-Da seems to be designed according to this principle. Her build has human elements but it is not ‘too real’: She can neither move, nor use her hands, her paintings are in fact executed by another artist, Suzie Emery11, her movements and answers are rather abrupt than smooth. Her visible artificiality serves as a slick rescue parachute sliding over the uncanny valley, making it easier to relate to her, or ultimately, from a marketing point of view, to take interest in her and the art she creates.
Taking a closer look at the art pieces she ‘makes’ (or more precisely codes since she can’t actually use her hands despite sketching), we can conclude that they are mostly inspired by an established canon of Western art. Most of her training data was based on pictures of artworks of already valued artists in the Western hemisphere like Pablo Picasso, Käthe Kollwitz, Max Beckmann, as well as writers Aldous Huxley and George Orwell. Her creations are mostly abstract or fragmentized to increase the perception of her being an autonomous artist. As Meller quotes himself: “We deliberately chose not to do a photorealistic view because people would just think it was something like printing when in actual fact, she is genuinely truly creative. The AI algorithm has a different outcome each time.”12We would however only scratch on the surface if we actually believe that Ai-Da is producing her own art style just because the outcomes are different. In the end, her artworks are human-machine collaborations, not only because it still depends on humans like Salaheldin Al Abd, Ziad Abass and Aidan Gomez, who constructed her Algorithms to create a human drawing style and to give her the ability to produce abstract artworks, but also because she is in fact not able to use her hands which makes her dependent on real-life artists like Suzie Emery (painting) or Alex Kafoussais (sculpture).
Teaming up AI with the art market appears to be very profitable: In 2018, the prestigious auction house Christie’s put up a piece produced by AI, an abstract, blurry portrait called Edmond de Belamy by the French art collective Obvious. It was estimated to bring in between $7,000 to $10,000. The piece was sold for $432,500, over 40 times more than expected.13Regarding the boom around AI generated art, it shouldn’t seem shocking that Ai-Da herself proves to be very profitable too: Her art in total already brought in $1 Million, establishing her creations as the new best thing collectors can buy into.14She also had her first exhibition ‘Unsecured Futures’, displayed in Lady Margaret Hall and St John’ College, University of Oxford in June/July 2019. The show was sold out – which seems hardly surprising: Ai-Da is presented as a well-planned marketing stunt by people working in the art economy (her creators are a curator and a gallery director), ringing every buzzword possible recalling technological progress and success on its way.
AI making art isn’t a completely new concept though. The first attempts of creating AI generated art already happened in the 1970s with Harold Cohens program AARON or, in the present time, with Google’s Deep Dream.15But Ai-Da proves to be by far the most successful of them all. Why so? Is it really the artistic brilliance of her artworks or isn’t her physical appearance the real novelty that arouses attention? Isn’t she herself the artwork that we buy into? Of course, with AI we can ask interesting questions like what we consider to be art and what it means to be creative, despite enabling new forms of art and creation. It’s certainly not my attempt to discredit AI-generated art with this question, but to shift focus on the gendered package it comes with.
Ai-Da’s creators’ team claim to have built Ai-Da as a chivalry to women in the art scene as they say they wanted to increase the representation of women in the field.16Although women are a powerful force in the art industry, their works aren’t sold nearly as much as the ones created by men. As Taylor Whitten Brown claims: “Only two works by women have ever broken into the top 100 auction sales for paintings, despite women being the subject matter for approximately half of the top 25.”17Despite any well-intentioned claims to boost the representation of women in the art scene, Ai-Da seems to be more of a halfhearted attempt to question gender equality. Other than sometimes expected, the art scene is still a club ruled by white, elitist men creating sexy femmebots, who produce uncountable pieces of art that are disposable for a juicy amount of money.
Design is never a coincidence, especially if it gets humanized and gendered. As Julie Carpenter et. al. state: “Humans often assign gender to inanimate objects; for example, cars and boats are commonly female in Western culture, and dolls are typically designed to be clearly male or female.” 18Looking at Ai-Da and other feminized humanoid robots like Sophia19, we can quickly see that a distinct higher percentage of AI robots are feminized by looks, voice and mannerisms. In fact, there is an epidemic regarding feminized technology: More and more anthropomorphized devices are gendered, like for example chat bots or digital assistants like Siri, Cortana and Alexa, getting sexually harassed by their users.20All of them programmed to behave obediently and serving, unable to talk back or to quit their jobs. As false copies of reality, of something that doesn’t really exist, or as Jean Baudrillard would say as a simulacra21, they represent womanhood from a patriarchal, heteronormative and binary perspective with their big eyes, small noses, full lips and soft facial features.
Ai-Da is no exception to that. She mirrors an established artist: not too young, elegant, feminine and beautiful – a fetishized object of desire, feeding into the romanticized fantasy of a sensual female artist. Following feminist theories on the performativity of gender and its social and ideological construction as – amongst others – Judith Butler postulates, we can understand the depicted femininity to be more of an operating system, unmasking gender as a performance.22Moreover, we can argue that this on-going feminization of technology roots in the notion that women always mark ‘the Other’ like Simone de Beauvoir claims in The Second Sex: “Humanity is male and man defines woman not in herself, but as relative to him; she is not regarded as an autonomous being. (…) He is the Subject, he is the Absolute – she is the Other.”23This Otherness understands womanhood to be something made or as De Beauvoir states herself: “One is not born, but rather becomes, a woman.”24And by becoming, or being made into a woman, we can draw the conclusion why so many AI robots are feminized. Technology is made by humans, but within the still very male-dominated field of IT, it reveals itself to be man-made. Ai-Da is yet again no exception to that, as her creators are mostly white, elitist Oxford men. Establishing gender as a marketing strategy ultimately reinforces the thesis that not only sex sells, binary and misogynist constructions of gender do too.
The feminization of technology as Helen Hester coins it,25has a huge impact on how we understand gender and eventually how we value women. While on the one hand patriarchal and ideological constructions of femininity (mostly by men) influence how such technologies are built, they are, on the other hand, also influencing how we see women and characterize them. This interaction results in real-life consequences for women all over the globe since they are, for instance, being heard less due to feminized voices in public spaces like train stations or airports.26Or as being taken less seriously since their technological simulacra are never questioning their circumstances or lacking ability to stand up for themselves. By creating feminized AI robots to boost equal gender representation in the art scene, they are obscuring the fact that women are still systematically discriminated against. Binary and patriarchal ideas of femininity seem like something they can craft and obtain, ignoring the fact that femininity is experienced differently and does not reveal itself as one absolutistic identity.
The demons we are apparently summoning are mostly beautiful, feminized AI robots, having a negative impact on women worldwide, demonizing them if they don’t act like their technological replicas. And the link between demonizing technology and women is, in fact, a strong and old one, as Glenda Shaw-Garlock claims: “As soon as technology became perceived as ‘a demonic, inexplicable threat and as harbinger of chaos and destruction’ it was reimagined as a woman.”27Demonizing women, and with Ai-Da especially women in the art scene, we must ask: Who is summoning this said demon? Who is it serving in the end? And who benefits from it? We can conclude that neither Ai-Da nor women in the art scene really benefit from this – the ones who do, not surprisingly, are the male puppeteers behind the curtains of such AI robots. With embodied AI, we are summoning the she-mon and the Armageddon she brings upon us will for sure hit very differently for women than it will for men.
1 McFarland, Matt, “Elon Musk: ‘With artificial intelligence we are summoning the demon’”, The Washington Post, https://www.washingtonpost.com/news/innovations/wp/2014/10/24/elon-musk-with-artificial-intelligence-we-are-summoning-the-demon/, published 24th of October 2014, accessed 22nd of January 2020.
2 To be read here: https://www.greekmyths-greekmythology.com/myth-of-pygmalion-and-galatea/.
3 To be read here: https://www.jewishvirtuallibrary.org/the-golem.
4 „Metropolis“, Fritz Lang, published 1927.
5„Westworld“, Jonathan Nolan/ Lisa Joy, published 2016 - , HBO; “Ex Machina”, Alex Garland, published 2015.
6 Manovich, Lev, AI Aesthetics, Moscow: STRELKA 2019, p. 8.
7 Knight, Will, „The Dark Secret at the Heart of AI, MIT Technology Review, https://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-ai/, published 11th of April 2017, accessed 22nd of January 2020.
10See Mori, Masahiro, “The uncanny valley”, IEEE Robotics and Automation, 19/2 2012, p. 98-100.
12“ADEBANJI ALADE DISCOVERS AIDA, THE FIRST ULTRA REALISTIC HUMANOID AI ROBOT ARTIST.”, Addictive Sketcher, https://www.youtube.com/watch?v=iqr2RN3ZzI4, published 22.06.2019, accessed 22.01.2020, 02:40-02:55.
13Cohn, Gabe, “AI Art at Christie’s Sells for $432,500”, The New York Times, https://www.nytimes.com/2018/10/25/arts/design/ai-art-sold-christies.html, published 25th of October 2018, accessed 22nd of January 2020.
14 Rea, Naomi, “A Gallery Has Sold More Than $1 Million in Art Made by an Android, But Collectors Are Buying Into a Sexist Fantasy”, artnet news, https://news.artnet.com/opinion/artificial-intelligence-robot-artist-ai-da-1566580, published 06th of June 2019, accessed 22nd of January 2020.
16 Rea, Naomi, “A Gallery Has Sold More Than $1 Million in Art Made by an Android, But Collectors Are Buying Into a Sexist Fantasy”.
17 Whitten Brown, Taylor, “Why Is Work by Female Artists Still Valued Less Than Work by Male Artists?”, artsy.net, https://www.artsy.net/article/artsy-editorial-work-female-artists-valued-work-male-artists, published 08th of March 2019, accessed 22nd of January 2020.
18 Carpenter, Julie et. al., “Gender Representation and Humanoid Robots Designed for Domestic Use”, in International Journal of Social Robotics 2009, p. 261 – 265, here p. 261.
19 For everyone interested: https://www.hansonrobotics.com/sophia/
20 Hester, Helen, „Technically Female: Women, Machines, and Hyperemployment“, SALVAGE, https://salvage.zone/in-print/technically-female-women-machines-and-hyperemployment/, published 08th of August 2016, accessed 22nd of January 2020.
21 Baudrillard, Jean, Simulacra and Simulation, Michigan: University of Michigan 1983.
22 Like in Butler, Judith, Gender Trouble: Feminism and the Subversion of Identity, London: Routledge 1990.
23 De Beauvoir, Simone, The Second Sex, London: Everyman 1993, p. 3.
24 Ibid. p. 283.
25 Hester, Helen, „Technically Female: Women, Machines, and Hyperemployment“.
26 Power, Nina, “The dystopian technology of the female voice”, http://hernoise.org/nina-power/ published 2012, accessed 22nd of January 2020.
27 Shaw-Garlock, Glenda, Descartes’ Daughters: Monstrous Machine-Women Through Time, Ottawa: Simon Frasier University 2006, p. 14.