The Algorithmic Gaze: Representations of Women in AI Art

Employing John Berger's Ways of Seeing (1974) as a critical lens, Danielle King observes the rise of the algorithmic gaze—the perpetuation of the male gaze into the realm of AI-generated art—underscoring entrenched patriarchal values and the protracted commodification of female bodies.
About the Author
Copper Frances Giloth, Modeling the Female Body – A Survey of Computer Generated Women 1980-1993 (Still), 1994

The Algorithmic Gaze: Representations of Women in AI Art

Employing John Berger's Ways of Seeing (1974) as a critical lens, Danielle King observes the rise of the algorithmic gaze—the perpetuation of the male gaze into the realm of AI-generated art—underscoring entrenched patriarchal values and the protracted commodification of female bodies.

John Berger famously articulated that "Men look at women. Women watch themselves being looked at" in his seminal BBC series Ways of Seeing (1974). This profound observation elucidates the power relations inherent in the act of viewing, wherein women are often objectified and positioned as passive subjects of male desire—a dynamic known as the male gaze. Reducing women to objects of scrutiny—deeply embedded in traditional art forms—has found new expressions in the realm of AI art, where algorithmic biases and societal conditioning continue to inform women’s representation. Employing Berger as a critical lens, we worryingly observe the rise of the algorithmic gaze—the perpetuation of the male gaze into the realm of AI-generated art—underscoring entrenched patriarchal values and the protracted commodification of female bodies.

The Male Gaze

British film theorist Laura Mulvey coined the term fifty years ago, writing in her 1973 essay “Visual Pleasure and Narrative Cinema” that the “male gaze projects its fantasy onto the female figure, which is styled accordingly.” Yet how has the male gaze remained such an enduring phenomenon? We look to science for clues. A 2012 study published in the European Journal of Social Psychology titled “Seeing women as objects: The sexual body part recognition bias” suggests that two distinct cognitive processes are at play with our perceptions of men and women. A summary of the study in Science Daily explained: “Participants in the study processed images of men and women in very different ways. When presented with images of men, perceivers tended to rely more on "global" cognitive processing, the mental method in which a person is perceived as a whole. Images of women were more often the subject of ‘local’ cognitive processing, or the objectifying perception of something as an assemblage of its various parts.”

Sarah Gervais, the study's lead author, explains, “Local processing underlies the way we think about objects: houses, cars and so on. We don't break people down to their parts -- except when it comes to women. Women were perceived in the same ways that objects are viewed."

This objectification of women extends centuries in art history and transcends media, with traces throughout digital art’s history, including its very beginnings. “The most widely circulated early artwork made using a computer,” according to Buffalo AKG, is a bitmap mosaic of a reclining nude woman titled Computer Nude (Studies in Perception I). This 1967 work by Kenneth Knowlton and Leon Harmon at Bell Labs reminds of Titian’s Venus of Urbino (1538), which—at over four hundred years old—remained the female nude prototype for centuries.

Titian, Venus of Urbino, 1538

Leon Harmon and Ken Knowlton, Computer Nude (Studies in Perception I), 1967. Courtesy of the artists' estates and Buffalo AKG

Further, in 1994, artist and professor Copper Frances Giloth created the video piece Modeling the Female Body - A Survey of Computer Generated Women 1980-1993 which spotlights the historical trajectory of women's representation in digital art. Through a compilation of computer-generated imagery from decades of SIGGRAPH Video Reviews, Giloth highlights the evolution of stereotypes in digital representations of women. The piece serves as a critical archive, documenting how early digital art often replicated the same gendered biases found in traditional media. From hypersexualized avatars to idealized bodies conforming to narrow beauty standards, the video lays bare the complexities of digital femininity and its impact on societal perceptions. As Giloth stated nearly three decades later: “Now in 2021, I could do a new collection from the last 28 years, but I won’t because I don’t think the content would be much different.” Giloth’s work begs the question about the current state of women’s representation in digital art.

Copper Frances Giloth, Modeling the Female Body – A Survey of Computer Generated Women 1980-1993 (Stills), 1994.

The Algorithmic Gaze

In AI art, the male gaze has been perpetuated in what we term the algorithmic gaze, manifested through the depiction of women in ways that cater to stereotypical notions of beauty, sexuality, passivity and subservience. AI algorithms, trained on datasets that reflect historical and contemporary biases, often perpetuate these norms, resulting in artworks that reinforce existing power structures. When these datasets are laden with historical art, media representations and societal norms that have long objectified women, the resulting AI-generated images often perpetuate this objectification, which may further amplify these biases.

Aesthetically, AI-generated artworks tend to reflect these biases through the portrayal of women in idealized, often sexualized forms. The works frequently highlight conventional beauty standards—young faces, slim figures, flawless skin, suggestive poses—that please the male gaze. This aesthetic preference is not merely incidental but deeply embedded in the way AI interprets and generates images based on its training.

The impact of the male gaze in AI art is not only aesthetic and conceptual but also economic. Sales data and market trends indicate a marked preference for artworks that conform to traditional portrayals of women.

AI-generated artworks featuring hyper-sexualized or unrealistically perfect female forms tend to fetch higher prices and receive more visibility in the digital art market, where the majority of collectors are men. This preference underscores the entrenchment of patriarchal values within the art market, where the commodification of female bodies continues unabated.

Challenging Norms

There are many notable efforts by artists, curators, scholars and technologists to subvert these entrenched, biased narratives and challenge the dominant paradigms of representation.

The exhibition In/Visible, curated by Senegalese artist Linda Dounia, showcases works by ten Black artists using AI today. These AI works interrogate the notions of visibility and invisibility in relation to gender, race and identity, directly commenting on the phenomenon of the biases present in AI art. As Dounia states in her Curator’s Note: “Any Black person using AI today can confidently attest that it doesn’t actually know them, that its conceptualization of their reality is a fragmentary, perhaps even violent, picture.” The exhibition includes works that critically engage with the portrayal of women. The Black women artists in In/Visible offer alternative perspectives that emphasize agency, complexity and diversity, disrupting the male gaze. Minne Atairu’s Blonde Braids Study II, for example, examines “the ways in which a text-to-image algorithm—Midjourney (v4)—renders a portrait of Black identical twins adorned with blonde braids. The resulting image underscores significant gaps in the training data, which inevitably precipitates a flattened representation of the Black identity outlined in the text prompt.”

Experience Minne Atairu's Blonde Braids Study II (2023) on ...
Minne Atairu, Blonde Braids Study II, 2023. Courtesy of the artist and Feral File

Artist Jake Elwes uses AI to explore themes of gender fluidity and non-binary identities, challenging the binary and often reductive representations of gender. Elwes’s project Zizi - Queering the Dataset involves training AI on a dataset of drag performances, producing artworks that celebrate gender fluidity. This approach not only subverts the traditional male gaze but also expands the possibilities for how AI can be used to represent diverse identities.

The anonymous performance artist OONA recently debuted a video essay exploring the issue of the portrayal of women in art, specifically within the NFT art space. In THIS TECHNOLOGY IS OUTPACING OUR WAYS OF SEEING, OONA—also invoking Berger—spliced together clips from the series, replacing the 1970s-era advertisement images of women with nude women—each very much an idealized, sexualized subject of the male gaze. OONA juxtaposes these nude NFTs with classical oil paintings in order to  highlight the lack of progress regarding the way women are portrayed in art.

The market preference for unrealistic, almost cartoonish, depictions of women extends not just to young women but to the rarer instances of AI art depicting older women as well. In contrast to the stereotypical and comical AI portrayals of older women that have seen major NFT market success, Francien Krieg’s realistic, unflinching, AI-assisted depictions of aging women have largely gone under the radar. Krieg’s works show older—often nude—women, representing a gamut of body types with wrinkles, sagging and all that comes with age. Krieg is a classically trained painter who has in recent years begun to incorporate AI into her practice, using both methods of art-making to “portray women in their natural state, without the artificial enhancements commonly found in the media. This depiction of older women in their raw and unapologetic form is not only empowering but also an important step towards breaking the stigma surrounding aging and the female body.” It’s unfortunate but perhaps not surprising that a number of Krieg’s striking pieces remain unsold, while sexualized images of impossibly perfect, thin young women sell briskly despite higher price tags.

Additionally, the inclusion of feminist theory and critique in AI art practices is gaining traction. Kate Crawford—a leading scholar of artificial intelligence and its impacts—and Joanna Bryson—a Professor of Ethics and Technology—have emphasized the importance of addressing gender and other biases in AI development and deployment. The work of these AI scholars advocates for more inclusive and equitable practices in AI, which artists have increasingly adopted to challenge the status quo. By incorporating feminist critique, these artists are not only creating more diverse representations but also pushing for systemic changes in how AI is developed and utilized.

Moving Forward

The portrayal of women in AI art reflects a complex interplay of historical biases, societal norms and market dynamics. My observation of the “algorithmic gaze” reflects the dominant patriarchal values within the art market and mass media, where the commodification and hyper-sexualization of female bodies remain ongoing issues. It is not just an aesthetic issue, but one with broader social implications. While the male gaze continues to exert a significant influence, privileging works that adhere to traditional representations of femininity, there is a growing movement to challenge and subvert these norms. By acknowledging persistent historical shortcomings through education and awareness—and diversifying the datasets used to train AI models—we can reimagine the possibilities for representing women. Transcending objectification and commodification, we can present and accept women as the whole, varied and complex beings that they are.


Danielle King is an artist, collector, writer and curator based in Western Massachusetts. Her recent work has utilized AI technology to create alternative art histories, explore memory and the duality of self, and investigate capitalist and art historical ideals of beauty and femininity. After receiving her MBA from the Yale School of Management, King spent eight years managing the Department of Painting and Sculpture at the Museum of Modern Art in New York. King is currently the CFO & COO of ClubNFT and Right Click Save.