Between Collective Memory and Contemporary Media: AI in Daniah Alsaleh’s Creative Practice

In March 2023, images of Pope Francis in a Balenciaga puffer jacket took the internet by storm. The stylish papal puffer images, which many believed to be real, were generated using artificial intelligence software Midjourney and opened the floodgates of debate around deepfakes and the difficulty of discerning whether an image is AI-generated or authentic.

Deepfake technology, which has been around for less than a decade, is infamous for its malicious use in politics, entertainment, and business to imitate the likeness of individuals. However, some artists see generative AI as a tool no different from a paintbrush or camera and an opportunity to enhance existing modes of artistic production. Machine learning software, such as StyleGAN, Midjourney, and Runway ML, among many others, has been used by artists in the past few decades to explore the frontiers of art making.

In the Gulf, countries have rapidly begun to embrace AI across various industries, including the creative sector. In 2022, Art Dubai, one of the largest regional commercial art fairs, introduced a digital section focusing on nonfungible tokens, video art, virtual reality, and, more recently, AI-generated artwork. In Riyadh’s historic Diriyah district, the Diriyah Art Futures arts, research, and education center focuses on creative emerging media and technologies, offering residencies and training programs in these fields. Similarly, in April, the Outer Edge summit in Riyadh gathered experts to consider the implications of AI and decentralized Web3 tech – where users have greater control and more of a stake – and to explore the future of the internet, virtual reality, gaming, smart cities, and more.

Probing the Past

Saudi multimedia installation artist Daniah Alsaleh has been using AI and machine learning in her work for several years, exploring the potential of this tool to elicit emotional reactions in viewers and manipulate historical narratives. In her practice, Alsaleh balances the fine line of believability, often relying on deepfakes – and welcoming their distortion and technical glitches – to challenge the certainty of memory as time and the deluge of media make their way into our minds.

Still from “How Do I Remember Thee” (2020). (Courtesy of Daniah Alsaleh)

In her 2020 video work “How Do I Remember Thee,” Alsaleh compiled archival images of Iraqis from the 1940s to 1960s – notably, from Iraq’s “peaceful pre-conflict era” – and trained an algorithm to generate deepfakes based on these images using StyleGAN. Alsaleh’s distorted and altered voice reads aloud her own eponymous poem as an audial backdrop. Shown as part of her Master of Fine Arts show at Goldsmiths, University of London, where she earned a degree in computational arts, “How Do I Remember Thee” examines the precarious relationship of memory formation to contemporary media. With billions of images shared online daily, determining what is AI-generated visual content and what is not can be challenging. Furthermore, this inundation of media, as Alsaleh’s work shows, can influence memories around certain events. At her studio in Hayy Jameel, Alsaleh told AGSIW that many visitors to her exhibition believed they recognized the faces in her AI-generated portraits despite them being derivative.

“The images evoked strong feelings of eerie familiarity, as if they are witnessing memories that are both theirs but not theirs. These works tapped into the layers of nostalgia, personal history, and cultural identity, creating an emotionally charged experience. This blend of the familiar and the uncanny triggers some sort of subjective connection to the context of the artwork.” She continued, “When viewers struggle to differentiate between what is real and what is AI-generated, it mirrors the broader societal challenge of navigating a media-saturated world where truth and fiction are increasingly intertwined. This confusion compels viewers to question the authenticity of their own memories and the sources of the images and media they consume on a daily basis. This feeds directly into my practice, which explores social conditioning and its influence on memory, reflecting on how we often take things at face value.”

Still from “Evanesce” (2021). (Courtesy of Daniah Alsaleh)

Another video work by Alsaleh, “Evanesce,” similarly draws on the ability of deepfakes to create somewhat realistic visual content based on archival material. The product, a 1950s-esque Egyptian drama that evokes the “glamor, romance, and sensuality” of Arab cinema’s golden age, was created by feeding Egyptian films into an algorithm to generate a 10-minute version of these films (without a discernible plot). Like the protagonists in “How Do I Remember Thee,” the characters and scenes in “Evanesce” show clear signs of digital manipulation and generation: Their facial features move unnaturally; their body parts are out of proportion; and background elements, such as palm trees and cars, shapeshift into unrecognizable versions of themselves. Knowing that her audiences will recognize the motifs that comprise black-and-white Egyptian films, Alsaleh’s short film, like her other work, relies on collective memory to make sense of the otherwise absurd scenes.

Perhaps paradoxically, Alsaleh’s practice involves using today’s most advanced tech to reevaluate old archival materials, which become the content that feed her algorithms. “I think one of the key opportunities AI offers is the ability to interpret and reinterpret archival subjects, providing different speculations and alternative ways of looking at things. To create such works, it’s essential to build a database to train these AI programs. This involves a continuous process of searching, collecting, and categorizing data, which often opens up new insights and perspectives that never crossed my mind.”

Still from “Evanesce” (2021). (Courtesy of Daniah Alsaleh)

Much like how Egyptian dramas challenged social norms by openly displaying elaborate love stories and musical performances, Alsaleh examines how an AI-generated film of this style can challenge contemporary engagement with media and the (re)construction of memory. Her practice welcomes debates around the believability of AI and its ability to produce meaningful artwork, allowing her work to exist in this gray area. Drawing on the realism of visual content generated with AI, rather than being threatened by it, her work uses AI as a conduit for the exploration of memory, the mind, archival history, and imagery.

AI in the Artist’s Studio

While debates around the increasing ubiquitousness of AI center around workers’ fears of being replaced in the workplace and questions of ownership and authenticity, Alsaleh sees AI as an opportunity to expand her work. She explained, “The way I use AI is as a collaborator rather than a competitor. This collaboration enhances creativity rather than replacing it. AI is just another tool in the toolbox, alongside brushes, paint, and cameras. Let’s not forget that art is inherently human, rooted in individual and collective experiences. AI cannot replicate the human experience and sensitivity that drive artistic creation. The human artist guides the AI, whether through prompt engineering, input commands, or personalized datasets.”

She continued, “We’ve seen similar shifts before when new technologies or innovations transformed their fields, such as the advent of photography. Just as photography became a powerful tool for artistic expression, adding to rather than replacing traditional art forms, AI can expand the boundaries of imagination while preserving the essential human element of artistic vision.”

The views represented herein are the author's or speaker's own and do not necessarily reflect the views of AGSI, its staff, or its board of directors.