Grant use vocal fry as a strategic method to cultivate positive regard from her peers, so disarming them. Consequently, individuals tend to underestimate her personal qualities and aptitudes.
A videoessay on AI labor, authored by Dayna McLeod, drawing upon labor from an AI writer and AI voice actor. Additional commentary by Lisa Messeri.
Capitalism Machina is a video essay that pairs AI writer Alisor South's essay The Pleasure Panic with an excerpt from Ex Machina directed by Alex Garland (2014). I used Kaylee, an AI voice actor to narrate the essay and matched this voice over with an excerpt from the film featuring Alicia Vikander who portrays Ava, an AI robot who is more sentient and self-aware than her creator realizes. This video essay uses these representations and creative outputs of AI against the backdrop of cisheteropatriarchy as a means to illustrate the similarities between how we dismiss, devalue, and underestimate women, particularly young women, and Artificial Intelligence.
South’s essay discusses capitalism in relation to the doctrine of human capital as a way to view humans as property and justify the exploitation of their bodies. It also argues that systems like capitalism and neoliberalism are pervasive, but not determinant. Again, this essay is AI generated. The AI writer (meaning the writer is AI) has not been assigned a gender; their name and bio do not indicate a gender. Having experimented with different voices and seeing their gendered effects when combined with the film, I decided to use Kaylee by Replica Studios to make visible the complexities of female identity, its construction, and how it has been shaped through the lens of cisheteropatriarchy. As Kaylee speaks, Ava literally constructs herself onscreen. The character bio for Kaylee on the Replica Studios website describes her as “from the sun-soaked shores of California, Kaylee is all about the West Coast Spirit and vibe. A little bit of drama a whole lot of sass, this California girl is a great supporting starlet.” Kaylee has three settings: Angry, Distraught, and Sassy. Each setting features varying degrees of vocal fry, and I’ve used each of them in my video essay as a means to emphasize a range of emotional escalation and as a way to disrupt a potentially monotonous flow of narrated text. I was drawn to this voice because Kaylee sounds like a young woman who is underestimated and seems like a perfect choice to underscore the similar underestimation of Ava in Ex Machina. Kaylee’s voice also reminds me of Shalita Grant’s outstanding portrayal of rookie lawyer Cassidy Diamond in season 3 of Search Party (HBO Max). In this show, Grant uses vocal fry as a way to gain favor with those around her by disarming them, and consequently, people underestimate her character and her abilities.
The excerpt I’ve taken from the end of the visually rich Ex Machina is when—SPOILER ALERT—Ava makes her calm escape from the remote luxury compound after murdering her creator and locking up Caleb (played by Domhnall Gleeson), a computer programmer who has been tasked with testing Ava’s consciousness, capacity for original thought, and self-awareness. This section of the film immediately came to mind when I first read South’s essay because this scene appears to reflect and illustrate the AI writer’s critique of capitalism:
“Treating people as human capital is how we justify the exploitation of bodies for profit, whether those bodies are working in coal mines or laying microchips on an assembly line. Human capital is nothing more than a way to view humans as property. Capitalism is not simply an economic system; it's a logic that permeates every aspect of our lives.”
This section of the essay neatly corresponded with shots of Ava assembling the surface of her body by applying synthetic skin to herself that she’d just taken from other AI robots, and through her self-examination in the mirror as she appreciates and evaluates her work on herself. When I was editing the video essay, I let the film continue on my timeline and have not cut or reassembled the film: I’ve simply juxtaposed the narrated essay and film to create and reveal meaning from these sources about female agency and the power of AI, and the consequences of messing with or trying to control both.
Ex Machina, dir. Alex Garland, 2014.
Voice over by AI actor Kaylee, Replica Studios [www.replicastudios.com]
South, Alisor [AI writer]. The Pleasure Panic. heliotropejournal.net, Feb 9, 2022 [https://www.heliotropejournal.net/helio/the-pleasure-panic]
Blake, Meredith. How ‘Search Party’s’ Shalita Grant turned her heartache into TV’s funniest millennial. Los Angeles Times, July 21, 2020 [https://www.latimes.com/entertainment-arts/tv/story/2020-07-21/shalita-grant-search-party-cassidy-diamond]
Lisa Messeri, Yale University
What are the questions that we, as critical scholars, ought to be asking about the onslaught of what are being called generative AI tools (ChatGPT, Dall-E, etc)? Dayna McLeod ends her video essay with three questions, “Does it pass? Does it fail? Does it matter?” The “it” in these questions tracks across the layers of the piece: “it” is Eva, the robot character in Ex Machina; “it” is Alisor South, the AI author who provides the text for the video essay; “it” is Kaylee, the AI voice actor who narrates the piece; “it” is the video essay itself. One of McLeod’s interests is drawing attention to those who are overlooked and underestimated, suggesting a synergy between women who look and sound a particular way (embodied and expressed by Eva and Kaylee) and these emergent AI creative assistants. The pervasive gendering of AI assistants (Phan 2017) are figuring into these next generation tools, with heightened normative consequences. Kaylee, for example, is only allowed to express anger, distress, or sass; Eva is assumed to be forever subservient. We accept these assumptions at our own peril, the video essay warns.
This piece opens up additional, implicit queries when thinking about the “it” of the video essay and whether “it” passes or fails. How does it “work” as a piece of scholarship? To put all of these tools together (even if in an admittedly “on the nose” assemblage) is indeed a contribution. These layers work to draw out McLeod’s analytic foci on gender, agency, and capitalism. But for this aspect of the essay to work, the dimensions of work that went in to producing the AI tools are left unexamined. It behooves us to both experiment with these generative AI tools while also bringing an analytic skepticism to how we are being told to work with these tools.
McCleod presents us with the categories of “AI writer” and “AI actor,” but I found myself uncomfortable accepting these titles. If the AI is a writer, then we are to conclude that it is they who are critiquing capitalism. But without knowing the humans, algorithms, and data sets that generated this text, the assignation of critique feels inappropriate. As our collective literacy of AI grows, we are better able to understand that these generated texts are predictive, creating best guesses based on a human inputted prompt as to a plausible sounding response. AI researcher Dan McQuillan has called these technologies “bullshit generators” and countless examples exist of factually incorrect responses from generative AI chatbots. Likewise, the company that produced Kaylee calls her an “AI actor” but here again I am unwilling to make equal the creative work that goes into voice acting with the engineering and algorithmic work that transforms an actor’s pre-recorded utterances into whatever the script prompts.
The questions of whether “it” passes or fails is less pressing than McLeod’s third question of whether “it” matters. To ask if it passes or fails accepts the Turing Test as the approach for assessing if these AI tools work. But to ask if it matters is to open up an opportunity for rejecting the categories and frameworks of the AI innovators and instead bringing in the analytic perspectives from the humanities and social scientists, better suited for questions about how and why AI matters. As McLeod deftly shows, generative AI can be used to probe thematic concerns that have long been central to our disciplines. And these same concerns can be productively directed back onto the work that goes into producing these tools, teasing out the different strategies by which AI is made to matter.
Phan, T. (2017). The Materiality of the Digital and the Gendered Voice of Siri. Transformations (29): 23-33.