To speak in the same breath of artificial intelligence and Asians, who have long been coded in popular culture as task-driven robots and emotionless cyborgs, risks belaboring the stereotype or, at best, rehashing well-worn critiques of techno-orientalism. Yet, four contemporary AAPI and Asian diaspora artists are charting new territory for this discourse with recent work that both embraces and deploys digital tools in experimental and affective ways, shattering stereotypes that encumber both humans and technology.
In the 60th installment of the Venice Biennale, “Stranieri Ovunque – Foreigners Everywhere,” New York–based multimedia artist WangShui presented Cathexis I, II, and III, a 2024 series of etched aluminum panels that, when nestled inside the arched windows of the Arsenale, resemble translucent alabaster haloed in natural light. On the panels, WangShui has hand-etched long, lithe serpents, which they position as a metaphor for data.
“When you zoom in on the serpent, its scales become granular data that is then multiplied across the surface to create a pattern that tessellates into a form,” they told ARTnews. “We are constantly zooming in and out of the data that drives our culture in order to understand it on a granular level and understand the superstructures it creates.”
The serpent reappears in the artist’s other work at the Biennale, Lipid Muse (2024), a long horizontal sculpture comprising thousands of LED lights pulsing in accordance with a multichannel simulation running in the backend, like a sloughed-off snakeskin with a life of its own.
Collaborating with programmers Brandon Roots and Caco Peguero on Lipid Muse, WangShui used machine learning to simulate what the artist called “quantum love,” whose frequency and vibration could devour forces like violence and repression. In the Arsenale with WangShui’s works, one encounters not the figure of a lover but a sensation as though the beloved had just stepped out, leaving dazzling impressions on the room in their wake.
At the Biennale, WangShui did not wish to depict a human body in a fixed form. They’ve always found ways to visualize states of transition and multiplicity. For instance, in “Window of Tolerance,” the artist’s first European solo show, which opened at Haus der Kunst last year, WangShui presented a multichannel live simulation titled Certainty of the Flesh (2023). Projected across LED panels, this work inhabited the format of a reality TV program in which AI avatars with translucent metallic skin struggled, in the artist’s words, to “take control” of their physicality and, as a result, ended up “breaking the form of the human body.”
Experimenting with neural networks is one way WangShui short-circuits the logic of identity categories, which, they noted, “end up trapping us.” For WangShui, who was born in Texas and lived in Thailand with their missionary parents as a child, the pressure to perform in accordance with others’ assumptions and projections is linked to the Freudian ego in the Western psychoanalytic tradition. Artificial intelligence allows WangShui to work toward a broader aim of “circumventing the ego” and finding a means of self-representation that does not require such performance. (Fittingly, the term “cathexis” refers to another Freudian idea, that of obsessing over an external object to expend libidinal desire.)
Another state of mind addressed by Freud—melancholia, or unconscious grief—was taken up as a framework by cultural theorist Anne Anlin Cheng and by critic David L. Eng and psychotherapist Shinhee Han in their case studies of Asian Americans in the mid-’90s. In a coauthored book, Eng and Han describe their research subjects as melancholic, citing their inability to let go of an unattainable ideal of whiteness. The melancholic subject, in this formulation, incorporates the unattainable object into their ego and sustains an ambivalent identification with it.
New York–based multidisciplinary artist and researcher Mo Kong explored such racial melancholia in their exhibition “Swift Island Chain” at the Brooklyn nonprofit arts organization Smack Mellon this past spring. Here, three cubicle-shaped sculptures were outfitted with desk dividers laser-cut with pithy auguries like EVERY YEAR / NEW SWIFT / DRIFTING IN and placed in the dimly lit exhibition space among vibrant text-based works on paper mounted on the walls and smaller totems made from biotech cement that contain the excrement of birds.
As in WangShui’s Biennale exhibition, Kong’s eerie post-pandemic office environment contained no fleshy bodies. “I like to include traces of human activity instead of representations of figures in my work,” Kong told ARTnews.
Instead, their exhibition featured three types of AI systems. The first—developed with data engineer Menyu Chen—translates lines of classical Chinese poetry into English three ways, to call attention to the linguistic nuances that are lost in transit between the languages. The translations are inscribed in cursive on the three “cubicles.”
The second system was designed to give feedback scores to the translations. “Every year / New swift / Drifting in” received a score of 081, according to the digital clock-font numerals etched on its respective desk. Connected to a Reddit forum made up of Asian immigrant workers, it represented Kong’s attempts to train their AI’s emotional capabilities: Kong exposed the system to the diction and syntax of disgruntled, displaced workers and, in doing so, drew a comparison between the rhetorical subtleties of web-based personal testimony and those of classical Chinese verse. In the third iteration, Kong mimicked the training of an AI in a series of wall-mounted works on paper showing their handwriting. Kong said they did so to show the repetitive manual labor of AI training, a welcome foil to the mechanical laser-cut text on the cubicles.
“The thing about AI labor is, it’s fast. It gives you outcomes immediately, and you come to expect that speed without seeing the calculations in the background,” they said during our interview.
During a studio visit in February, Kong pointed out that nascent forms of AI are already inheriting the stereotypes placed on Asian immigrant workers in American society, who have been stereotyped as suited to perform repetitive labor cheaply, efficiently, and without emotion. Kong disrupts this pernicious cycle of racial melancholia by cultivating an AI conversant with non-Western literature and emotional expressivity, and by working, with each iteration, to make conscious the loss accrued in the transition between languages and cultural traditions.
In April, on the final weekend that “Swift Island Chain” was on view, two New York–based artists, Tianyi Sun and Fiel Guhit, activated Kong’s installation with an AI-based performance titled Workloop (2024). Sun and Guhit, who collaborate on sound- and performance-based projects, set up computers, monitors, and chairs in Kong’s laser-cut cubicles. Over a jazzy piano score and clad in black boots, button-ups, and blazers, the pair role-played in dialogue with an autoregressive text-to-speech system that they built, speaking in real time to the system, which responded by drawing from conversations it was previously trained on.
In a sense mimicking the way their system was constantly running, the performers moved continuously throughout the performance, traversing the exhibition space, brushing past one another, leaning on columns, sometimes sitting, but not for long. Viewers lost track of how many times they performed the same or similar actions.
“So much of labor comes through repetition,” Sun told ARTnews after the performance. It is through repetition, after all, that workplace norms and systems of power are established.
When the two artists first started working together, Sun and Guhit were interested in temporalities and imaging technologies. Around two years ago, they recognized that AI could be used as a critical lens to describe how the contemporary subject experiences time. During previous performances, including one this past February titled Warmer Layers (2024) at the now defunct Helena Anrather Gallery in New York, Sun and Guhit used handheld microphones to capture errant noises over the course of an evening, including the babbling of a baby in the audience and the sound of the building’s elevator in operation. The AI then read the sounds as spectrograms and produced, as a live-generated composite of the night’s sounds, a series of decontextualized pronouncements akin to the parody of a manifesto mixed with a nursery tale: “I am the shield against the burning rays. I am the armor that protects your delicate skin […] I am your sunscreen, your sunblock, you sun-blocker. There once was a toaster who loved nothing more than making toast.” Later on, Guhit explained, “We wanted people to experience that moment of generation.”
Workloop, meanwhile, critiqued what Kong previously described as our demand for immediate outcomes. During one sequence, Sun role-played a call center bot whom the AI was calling about relatively minor dissatisfactions: the red wine it’d purchased wasn’t bold enough, a toaster was taking too long to toast bread. As Sun repeatedly “transferred” the AI, the program grew increasingly angry. “What kind of joke is this?” it eventually said. “Your customer service skills are abysmal.” Ironically, because Sun spoke to the AI and read poetry to it continuously during the training process, taking care that the program captured her tones, inflections, and cadences, the program’s voice sounded uncannily like hers.
Technology, Sun said, is perceived as needing to be “smooth, perfect, and seamless.” Here, echoes of Kong’s observations about Asian workers can be heard. “That was why it was important for us to build our own model and work on something that was, I would say, imperfect,” Sun added.
Guhit agreed, adding that it is often difficult to capture and utilize outputs they find desirable. Just when they think they have their program figured out, he said, “something else takes us by surprise.”
In this way, the performers’ willingness to improvise with AI and to be surprised by its anomalies resembles a kind of empathy that extends beyond our relationship with technology. For Sun and Guhit, this attitude comes as naturally as working with any other artistic medium.
“In the same way that, if we were to be using clay or paint as a live material during a performance, if something splashes, we work with it,” Guhit said.