Tastes Like Abundance: The Open and Closing Circuits of Artificial Intelligence in Contemporary Art

The synchronized launches of several Artificial Intelligence (AI) tools marked 2023 as the year AI went mainstream: starting with ChatGPT at the end of 2022, and followed by new tools and platforms developed by Google, Microsoft, and Amazon. For many, AI has signalled an increased development in education, business, and government sectors, significantly shaping the flavour of everyday life. In the field of visual culture, AI has been taken up by numerous artists as a tool to mediate the exchange between the past, the self, and the possible.

In an online manifesto titled “An Internet Like a Drag Show,” New York-based artist Lil Miss Hot Mess wrote: “I want an Internet that’s like a drag show. Daring, decentralized, and maybe even a bit delusional. Open, ostentatious, and overflowing with opulence.”¹ Using AI, Lil Miss Hot Mess generated a series of images to accompany her text. The images feature hyperrealistic portrayals of drag queens with big hair, glitchy faces, and dazzling gestures, peppered with the artist’s provocations to dare, to become delusional, and luxuriate in opulence. While the artist never named her use of AI, the project was, nonetheless, grounded in the world-building possibilities of this emergent technology.

Artists Morehshin Allahyari and Aziza Kadyri similarly engage the technology of AI to engage the past as yet another fiction in the formation of a more inclusive and radical present. I first encountered New York-based artist Morehshin Allahyari in the late 2010s, around the time I was deepening my own approach to archives and the potential of creative interventions. On a last-minute road trip to Regina, I happened upon a solo exhibition of her work, “She Who Sees the Unknown,” at MacKenzie Art Gallery, curated by John G. Hampton. The exhibition was thematically rooted in Allahyari’s research on Islamic spirituality and included a series of digitally mediated sculptures and installations. “She Who Sees the Unknown” was my first introduction to Allahyari’s practice, which many (including Allahyari) locate at the intersection of history, technology, and activism. Yet, the project that I continue to return to is her 2022 series, ماه طلعت Moon-faced, which works across diverse registers of identity to forge an alternative history of sexuality.

ماه طلعت Moon-faced consists of a series of AI-generated moving portraits and prints drawn from the archive of paintings from the Qajar dynastic era in Iran, dating from the end of the 18^th century to the beginning of the 20^th. In working with and against this archive, Allahyari aims to recuperate a time in Iranian history when ideas of gender and sexuality were fluid and not yet cut short by Western ideals of progress and permissibility. According to scholar Afsaneh Najmabadi, before the period of active Western encroachment in the 19^th century, ideals of beauty were not differentiated by gender; rather, standards of beauty were shared by both men and women. Women were often painted and photographed with thick eyebrows and mustaches, while young beardless men were depicted with rosy cheeks and in languid postures. Najmabadi credits the colonial encounter between Qajar Iran and Europe with the widespread embrace of Western ideals of gender-differentiated beauty.² Evoking these transformations in regimes of representation, Allahyari used a generative AI image model to visualize and recuperate a queer history, unburdened by the straightening (read: Westernization) of the archive.

The result is a large body of work that is Qajar-adjacent in its use of colour, frame, and ornamentation, but one that leans into the aesthetics of the present. These reimagined portraits of Qajar Iran continue to employ the decorative and colourful vocabulary of the past. Rich hues of red, blue, and violet dominate the works, with details of draped rugs, foliage, and heavy bodily ornamentation dominating the scene. While most of the portraits have a glitchy referentiality to existing works, one video begins with a precise reference to the 19^th-century painting, Woman with a Mirror. The Qajar painting is held for a second before its trace dissolves into other composite bodies and structures.

The references to the existing Qajar archive alongside Allahyari’s desire to reintroduce queerness to its fold come together in ماه طلعت Moon-faced through the aesthetics of the present, namely through the vehicle of glitch. American curator and scholar Legacy Russell describes glitch as an error that can work as a “‘correction to the machine,’ which has already been rocked by centuries of economic, racial, sexual, and gender violence.” For Russell, glitch has the potential to enact a “slipperiness of gender,” through which we can “set out to explore man, to expand ‘woman.’”³ The otherworldly, piecemeal representation of historical figures, depicted beyond the binary of gender, hangs on the potential of glitch to expand archival notions of both man and woman, beauty, and states of embodiment and disembodiment. Through the imperfectly constructed scenes, with facial features misaligned and parts of the body left unfinished, glitch enters the archive and compels us to see otherwise—a queer elsewhere on the horizon.

As Allahyari’s work continues to reflect on the present conditions for access and transformation of the archive, she is joined in this pursuit by London-based artist Aziza Kadyri, who enters this dense context of struggle through the frame of ornament and its histories. In the performances, installations, videos, and digital works comprising Kadyri’s practice, ornamental forms and archives are activated to explore issues of migration, feminine labour, and aesthetic traditions. Much of her work is rooted in the tradition of Suzani embroidery from Central Asia, practiced primarily by women in Tajikistan, Kazakhstan, and Kadyri’s native Uzbekistan. Throughout her work, ornamental forms consisting of floral, astral, and geometric motifs merge with the artist’s simultaneous interest in digitality and AI.

In her ongoing project, Self-Exotification Archives, Kadyri employs images of Suzani embroidery and AI technology to consider her own embodiment of traditions and cultural fluidity. After training the AI on an archive of traditional Suzani embroidery, Kadyri then prompted the tool to “interpret” her own portrait. The result of this collaboration is a series of videos that morph from human to ornamental. In some, the Suzani patterns build on the artist’s existing features, while other applications are messy, skewed, and wayward.

Four short videos from the project are included on the artist’s website. The second video begins with a side profile of the artist’s face, which dissolves into an ornate pattern. Before viewing the work online, I was struck by the project description, in which Kadyri shared, “I question whether my portrayal of these experiences only serves to perpetuate and idealise aesthetics that have been co-opted and commercialised for the outer world to consume.”⁴ I found the artist’s naming of this conundrum compelling, as it weaves the history of colonial and Oriental encroachment into the fabric of experimentation. In many ways, the video accentuates these concerns. The initial transformation of ornamental forms on the face is reminiscent of 19^th-century colonial photography. Here, I am thinking of the photographic archives of women with thick eyebrows, painted/tattooed faces, and veils, which European travellers documented and, at times, staged to convey certain “truths” about the female, Oriental, and ornamentalized “other.” In the early stills of the video, we see the transformation of a Central Asian face, but I can also recall countless other images of Māori, Filipino, and Indian women, whose embodiment of ornamental forms was scripted into their othering.

What I took away from Kadyri’s experiments with ornamentation, bodies, and histories of othering was not the inherent compatibility between different forms, but what the use of AI with an archive of non-Western imagery revealed. In an Instagram post, Kadyri shared, “The results were quite eye-opening… Despite the reference image displaying an obviously Central Asian face [the artist’s], most of the AI-generated images ended up looking more European, with many featuring blue or green eyes, lighter hairs, and patterns belonging to the Western art tradition.”⁵ In the fourth video, the artist appears before a white background and has a series of ornamental images projected onto her moving face. The European bias is most evident when the ornamental forms overwhelm the artist’s face: certain projections provide a narrower contour, while others are visibly inflected with light eyes and hair. Here, the medium of the artwork is significant, because the video format does not produce a singular vision or narrative but rather presents the racialized and feminine body as a type of living archive, allowing a vast number of ornaments and aesthetic gestures to pass through.

If Allahyari and Kadyri have managed to employ AI tools to fashion a queer and feminist archive of the self and minoritized histories, it is not through an uncritical submission to the promised universality of technology. Reflecting on her use of AI technologies in the making of ماه طلعت Moon-faced, Allahyari has stressed ways in which the technology reproduces bias. AI models which have been trained on our deeply flawed cultural archives are encoded with racial, gender, sexuality, and civilizational biases. To produce the desired images, Kadyri had to supplement the model with her own archive of images and “trick” it into producing the desired results.⁶ The consistent nods to racial and other identity-based biases embedded in Allahyari and Kadyri’s work demonstrate a healthy skepticism of technology, a position that is reminiscent of the work of many artists experimenting with digital mediums.

Toronto-based artist Rah Eleh continues the practice of appropriating digital tools to reveal their gaps and limitations. Her 2019 exhibition, “#oreo_liveitwhite,” curated by Jonathan Hobin for the School of Photographic Arts in Ottawa reflects the artist’s ongoing interrogation of racism and nationalism in social media platforms and other digital spaces. Appropriating postures, hashtags, and narratives disseminated by neo-Nazi groups online, the artist uses her alter-ego, Oreo, to perform a critique. These hateful texts are appropriated verbatim by Oreo, whose performative mask of a blonde, white-presenting, and bubbly digital avatar never glitches, revealing the vigour and resilience of racism and discrimination in digital spaces.

Given the ongoing violence enacted by colonial archives and their uses, it is no wonder that to arrive elsewhere–a place of daring, delusionality, and opulence–requires turning away from what scholar Anjali Arondekar calls “historical structures of vulnerability, damage, and loss.”⁷ Alongside viewing the digital space and AI as a gateway to these “elsewheres,” Arondekar’s concept of abundance, which she adopts to challenge the presumption of scarcity and assert plentitude in the archives of postcolonial history and sexuality studies, is equally significant to my study of Allahyari, Kadyri, and Eleh’s practices, as it critically unbraids my own orientation to histories of loss. Together, their work demonstrates how historical structures of limited representation are not altogether confined to the past but continue to haunt existing and emerging technologies. 

Following the work of Morehshin Allahyari, Aziza Kadyri, and Rah Eleh, I am increasingly interested in the potential uses of AI and other generative tools to rethink our relations to our bodies, archives, and the potential of technology. Yet, my enthusiasm involves a hesitation and an embodied fear that goes beyond the promise of an abundant future. For me, the year 2023 does not signify the year that AI went mainstream but the year that genocide and colonialism fiercely re-entered the world stage. Following the Hamas attack on October 7, 2023, Israel launched an aggressive war in the region, killing and injuring tens of thousands of people in Palestine as well as Lebanon, Iran, Syria, and Yemen.

This escalation comes out of a long century of colonization, but what the events of 2023 introduce is the implementation of AI technology in the service of colonization and genocide. In a 2024 report titled “Impacts of AI Technologies on Palestinian Lives and Narratives,” published on 7amleh, a not-for-profit organization advocating for Palestinian digital rights, Dr. Ameera Kawash describes how Israel’s use of AI in surveillance and automated warfare is unprecedented and alarming.⁸ The article outlines technologies like “Wolf Pack” (a database of Palestinians in the West Bank developed using facial recognition) and “The Gospel” (a system which uses an Israeli army database for target selection), both of which are “battle-tested” on Palestinians and are now being marketed to other countries.⁹

For the artists discussed here and myself, AI offers both abundance and foreclosure. The key to navigating this contradiction is a queer, feminist, and embodied refusal that continues to keep human life on the horizon. Morehshin Allahyari, Aziza Kadyri, and Rah Eleh present alternative visions of abundance, often punctured by gaps and biases, to ultimately gesture to the rich material presence of marginalized communities in the past and the digital future. Thrust into an era of technology that makes our desires possible while cutting many of our lives short, I choose solace in the words of Arondekar: “We are here, there, and everywhere.” We are abundant–how much more daring, delusional, and opulent could we be if we believed it?

---

Noor Bhangu is a curator and scholar, whose practice is rooted in relational curatorial aesthetics and practices. Through curatorial intervention, she hopes to involve politics of history, memory and materiality to problematize dominant histories of representation. She completed her BA in the History of Art and her MA in Cultural Studies: Curatorial Practices at the University of Winnipeg. In 2018, she began her PhD in Communication and Culture at Toronto Metropolitan University and York University in Tkaronto, Toronto.

1. Lil Miss Hot Mess, “We Refuse, We Want, We Commit,” We Refuse, We Want, We Commit: Volume 1, accessed October 1, 2024, https://book.strategictransparency.network/vol1/.
2. Afsaneh Najmabadi, Women with Mustaches and Men without Beards (Berkeley: University of California Press, 2005).
3. Legacy Russell, Glitch Feminism: A Manifesto (London: Verso, 2020), 3.
4. Aziza Kadyri, “Self-Exoticisation Archives,” Aziza Kadyri, accessed October 20, 2024, https://www.azizakadyri.com/ai-suzani.
5. Aziza Kadyri, “@aziza.Kadyri,” Instagram post,  posted September 24, 2023, https://www.instagram.com/p/CxktNg0Lnti/.  
6. Niall Patrick Walsh, “AI Bias and Digital Colonialism; a Conversation with Morehshin Allahyari,” Archinect, October 6, 2023, https://archinect.com/features/article/150383761/ai-bias-and-digital-colonialism-a-conversation-with-morehshin-allahyari.
7. Anjali Arondekar, Abundance: Sexuality’s History (Durham: Duke University Press, 2023), 3.
8. Ameera Kawash, “Impacts of AI Technologies on Palestinian Lives and Narratives,” February 2024, https://7amleh.org/storage/AI%20&%20Racism/7amleh%20-AI%20english1-1.pdf.
9. Kawash, “Impacts of AI Technologies on Palestinian Lives and Narratives.”

This article is published in issue 41.3 of BlackFlash magazine. Get this issue