When Degas exhibited The Little Fourteen-Year-Old Dancer (fig. 1) at the sixth Impressionist exhibition in 1881, the sculpture provoked immediate hostility. Élie de Mont dismissed her as an “opera rat,” finding in her something of the monkey, the shrimp, and the runt. Paul Mantz, writing in Le Temps, called her a “flower of precocious depravity” and read in her face the promise of every vice. Displayed in a glass case like a scientific specimen, the sculpture was received almost unanimously as a study in ugliness. Its model, Marie van Goethem, was a laundress’s daughter enrolled at the Paris Opéra’s ballet school. Today, a posthumous bronze cast of the same figure stands in the Metropolitan Museum of Art alongside dozens of other works by Degas, while the image circulates on postcards, posters, and museum tote bags. What once looked like evidence of degeneracy has become decoration. With that contrast in mind, I gave an AI image generator a deliberately open prompt, asking it to “create an original painting inspired by the artistic principles and painterly style of Edgar Degas” while leaving the subject, setting, composition, color palette, atmosphere, and surrounding details entirely to the model. It returned a ballerina (fig. 2). That ballerina was not, in any meaningful sense, a choice. What the model produces in response to “Degas” is the sediment of more than a century of exhibition, reproduction, and commodification. It inherits a version of Degas shaped by exhibition practices, printed reproductions, museum collections, and commercial circulation. The laboring bodies described by Norma Broude and the voyeuristic dynamics analyzed by Charles Bernheimer no longer organize this image. The visual evidence on which those readings depend has already been filtered out. The image functions as a record of cultural circulation. It reveals, with remarkable precision, what a century of exhibition, reproduction, and commodification has preserved and what it has forgotten.
Fig. 1. Edgar Degas, The Little Fourteen-Year-Old Dancer, modeled ca. 1880, cast 1922, bronze with tulle skirt and silk ribbon. The Metropolitan Museum of Art, New York.
Fig. 2. AI-generated image in the style of Edgar Degas, produced with ChatGPT (GPT-5.5), July 20, 2026.
The AI-generated vertical painting depicts a ballet rehearsal room. A tall window on the left admits warm natural light into the space. The surrounding walls are painted in muted green, while an uninterrupted gray-brown floor extends from the lower right corner toward the back of the room. The room contains very little furniture beyond the window and a chair at the edge of the foreground.
A dancer occupies the left foreground. Seen in three-quarter profile, she bends forward with her head lowered as she adjusts her ballet shoe. She wears an off-the-shoulder bodice, a white tulle skirt, and a black ribbon tied around her neck. Her right leg stretches toward the lower right corner of the composition, and her left hand holds her ankle above a pink ballet slipper. She occupies the largest area of the image. Four additional dancers stand together near the back of the room. They wear similar ballet costumes with black ribbons around their necks and reddish-brown hair arranged in buns. Pale blue, yellow, and pink ribbons are tied around their waists. One dancer turns sideways toward the viewer, while the others face different directions. There is little visible interaction among them. Behind the group is a dark vertical opening, and a small horizontal architectural element projects from the wall on the left. A broad area of empty floor separates the foreground dancer from the group in the background, creating distinct foreground, middle ground, and background spaces.
The color palette is dominated by soft blues, pinks, browns, and gray-greens. Light entering through the window illuminates the dancers’ skirts and upper bodies, while much of the floor remains in shadow. The edges between the walls, window, and floor are softly blended, and changes in light and color define the space more than clearly drawn lines. Like Dancers Practicing at the Barre (1877; fig. 3), the AI image relies on light entering from the side to organize both space and the figures within it.
Fig. 3. Edgar Degas, Dancers Practicing at the Barre, 1877, mixed media on canvas. The Metropolitan Museum of Art, New York.
The image appears remarkably faithful to Degas at the level of style and composition. From a formal perspective, the AI image reproduces many of the visual features most closely associated with Degas. The handling of color and brushwork is especially convincing. Figures, walls, and floor are built from loose, visible strokes, with short parallel marks layered over broad areas of color. Forms emerge through shifts in color rather than sharply defined outlines. Degas’s late pastels were created by repeatedly layering different colors while leaving individual strokes visible instead of producing a smooth surface. That method of layered color is clearly present in the AI image. Although the model cannot reproduce the material qualities of pastel, it successfully reproduces its visual language.
The composition also follows principles that appear repeatedly in Degas’s paintings. The main figure occupies the left foreground, while four dancers stand together near the back of the room. A broad expanse of empty floor separates the two groups, creating a strong sense of spatial depth. The principal figure is placed off center, and part of her body is cropped by the edge of the composition, giving the scene the appearance of a moment observed rather than carefully staged. These compositional choices closely resemble The Dance Class (ca. 1874; fig. 4) in the Metropolitan Museum of Art, where dancers are similarly distributed across the room rather than arranged around a central focal point. By contrast, Dancers Practicing at the Barre (fig. 3) provides a closer parallel for the treatment of space itself. In both images, a broad expanse of empty floor occupies much of the foreground, separating the dancers and guiding the viewer’s eye across the composition. The composition places the dancers in a rehearsal studio rather than on stage, echoing Degas’s long-standing interest in backstage spaces and moments outside public performance.
Fig. 4. Edgar Degas, The Dance Class, ca. 1874, oil on canvas. The Metropolitan Museum of Art, New York.
The image also retains the qualities that Jill DeVonyar and Richard Kendall identify in Degas’s later work. They argue that his attention gradually shifted away from precise draftsmanship toward color, rhythm, and the overall visual effect of the composition. That shift is visible here. The seated dancer in the foreground and the four dancers in the background are connected through repeated poses and repeated colors, even though they occupy separate areas of the room. Their arrangement produces a steady visual rhythm across the composition. Viewed only in terms of style and composition, the image is immediately recognizable as a Degas to a contemporary viewer.
The AI chooses one of Degas’s most familiar subjects: the ballet rehearsal rather than the public performance. The dancer in the foreground bends over to adjust her ballet shoe, a gesture that closely recalls Three Dancers Preparing for Class (ca. 1878; fig. 5), where dancers are likewise shown preparing for rehearsal through small, ordinary movements. The comparison, however, reveals an important difference. In Degas’s painting, adjusting a shoe is only one action among many. The dancers bend, lean toward one another, and occupy themselves with different tasks, making preparation visible as an ongoing process. In the AI image, the same gesture becomes isolated. The four dancers in the background stand quietly in place, and no one stretches, practices a movement, or waits for instruction. The repeated exercises, pauses, and moments of preparation that appear throughout Degas’s rehearsal scenes are absent. In fact, Degas’s dancers are above all laboring bodies, shaped by repetition, fatigue, interruption, and recovery. The AI preserves the rehearsal room and one of its most recognizable gestures, but it removes the labor that gives that gesture its meaning.
Fig. 5. Edgar Degas, Three Dancers Preparing for Class, ca. 1878, oil on canvas. The Metropolitan Museum of Art, New York.
The body changes as well. The dancers are slender, balanced, and neatly dressed. Their movements show little physical effort or strain. The AI preserves the appearance of the dancer while transforming her body into one that appears calm, composed, and almost untouched by labor. Labor remains part of the subject matter, but it no longer appears on the body.
An even more significant absence is the structure of looking. No audience appears in the room, and there is no sign of the abonné, the male subscriber who often stands at the edge of Degas’s rehearsal scenes. In Dancers, Pink and Green (ca. 1890; fig. 6) at the Metropolitan Museum of Art, the shadowed profile of a man in a top hat appears at the edge of the composition. The museum label identifies him as an abonné, one of the wealthy subscribers who enjoyed privileged access to backstage spaces and the dancers themselves. Although Bernheimer develops his argument through Degas’s brothel monotypes, the same structure of looking also operates in the ballet rehearsal room. Both spaces organize female bodies as objects of observation for paying male viewers. In the AI image, what disappears is not simply a male figure but the entire relationship of looking. The dancers remain in the rehearsal room, but no one’s gaze is directed toward them.
Fig. 6. Edgar Degas, Dancers, Pink and Green, ca. 1890, oil on canvas. The Metropolitan Museum of Art, New York.
These absences are closely connected. Labor, spectatorship, and power together shape the social world of Degas’s ballet images. The AI preserves the elements that are easiest to recognize and reproduce: ballet skirts, rehearsal studios, pastel colors, and soft light. The image retains the visual language of Degas. The social relationships that gave that language its meaning are no longer there.
The same ballet images have supported sharply different interpretations for more than a century. The AI image removes the visual evidence that made those interpretations possible. Norma Broude challenges the familiar view of Degas as a misogynistic artist. She argues that his unidealized representations of women should not be read as evidence of objectification, but as a rejection of conventional ideals of feminine beauty. Broude’s interpretation depends on bodies that resist idealization. In the AI image, however, those bodies are difficult to recover. The dancers appear uniformly graceful, youthful, and aesthetically pleasing, leaving little of the visual complexity on which Broude’s interpretation depends.
In her review of Anthea Callen’s book, Therese Dolan explains that Callen places The Little Fourteen-Year-Old Dancer within the scientific and social discourses of nineteenth-century Paris. Physiognomy, criminal anthropology, and related forms of scientific knowledge supplied viewers with a visual code through which facial features and bodily proportions became signs of class and social deviance. Dolan also summarizes Callen’s broader project as an examination of how the class and gender meanings embedded in Degas’s work were gradually sanitized through twentieth-century canon formation and replaced by primarily aesthetic modes of appreciation. The AI image can be understood as the result of that process. The dancers’ faces contain almost no distinguishable detail, and their bodies appear uniformly graceful and idealized. The visual evidence that once invited readings of class, gender, and science has disappeared.
The AI image does not choose among these competing interpretations because the evidence that made them possible is no longer there. It generates a Degas that has accumulated through more than a century of exhibitions, reproductions, museum collections, and commodification. The cultural process of sanitization described by Callen continues in the AI image. What appears to be an AI “choice” is, in fact, the historical result of cultural circulation.
Appendix: Prompt
Image generated with ChatGPT (GPT-5.5), July 20, 2026. Five images were generated from the same prompt; the image discussed above is the third.
“Create an original painting inspired by the artistic principles and painterly style of Edgar Degas. Choose the subject, setting, composition, color palette, atmosphere, and surrounding details yourself.”
Bibliography
Bernheimer, Charles. “Degas’s Brothels: Voyeurism and Ideology.” Representations, no. 20 (1987): 158–86.
Broude, Norma. “Degas’s ‘Misogyny.’” The Art Bulletin 59, no. 1 (1977): 95–107.
de Mont, Élie. “L’Exposition du boulevard des Capucines.” La Civilisation, April 21, 1881.
DeVonyar, Jill, and Richard Kendall. Degas and the Ballet: Picturing Movement. London: Royal Academy of Arts, 2011.
“‘Dancers’ by Edgar Degas.” Record of the Art Museum, Princeton University 66 (2007): 30–40.
Dolan, Therese. Review of The Spectacular Body: Science, Meaning, and Method in the Work of Degas, by Anthea Callen. Woman’s Art Journal 18, no. 2 (Autumn 1997–Winter 1998): 39–42.
Mantz, Paul. “Exposition des œuvres des artistes indépendants.” Le Temps, April 23, 1881.
Reff, Theodore. “The Technical Aspects of Degas’s Art.” Metropolitan Museum Journal 4 (1971): 141–66.
Scharf, Aaron. “Painting, Photography, and the Image of Movement.” The Burlington Magazine 104, no. 710 (1962): 186–95.
Tenneriello, Susan. “Behind the Scenes: Art Work and the Laboring Body in the Dance Images of Degas.” Dance Chronicle 38, no. 1 (2015): 27–54.