Recreating the Aura: Professor Naeem Mohaiemen Explores Artificial Intelligence and Art in the Classroom

In the ever-expanding discourse around art and AI, Associate Professor of Visual Arts Naeem Mohaiemen foregrounds curiosity.

Mohaiemen, who also heads the Photography concentration of the Visual Arts Program, recently developed a course in “Artificial Intelligence & Photography,” which he offered for two semesters. The course, which was supported by Google’s Artists + Machine Intelligence Grant and the Office of the Provost’s Junior Faculty Grants Program, explored the ethics and aesthetics of AI photography. 

A multidisciplinary artist, academic, and author, Mohaiemen channels his wide-ranging curiosities into a research-intensive practice. This approach to art-making translates to his teaching style. I asked, in our email correspondence, what compelled him to have his students experiment with AI.

The simple answer is that Mohaiemen received an invitation from Google to apply for their Artists + Machine Intelligence Grant, intended to facilitate collaboration at the intersection of art and technology. In his own work, Mohaiemen was interested in drawing attention to the scarcity of the archive—and thought that AI usage would conflict with that aspect of his studio practice. So, he applied for the educational version of the grant instead, and his class and Machine Visions lecture series were born of it. 

But the story is more complicated than that. Mohaiemen acknowledged that, like many of his colleagues, he feels that AI’s future impacts on the arts remain nebulous—the only certainty being the massive displacement it will bring to people working in the arts. Yet the fear surrounding the discourse is precisely what encouraged him to begin working with AI in the classroom. “The entire discourse is dominated by capital and industry,” he told me, “which has a vested interest in disruption and breakage, without thinking through the long-term consequences.” Fortunately, Mohaiemen has access to dozens of people who do have a vested interest in the long-term impacts of AI on creative cultural industries: art students. 

As any resourceful professor knows, the classroom is not only a space for imparting knowledge, but for experimentation and collaboration. Mohaiemen noted that he was hoping to learn from the class as well: “I was interested to enter this space so as to learn and develop a sharp critique, as well as to glean the positive aspects of this transformation. Teaching a class on this subject seemed the best way to train myself, to try things out, and to learn from and with my students.”

The class considered AI within the long genealogy of machine visual intelligence, beginning with Walter Benjamin’s seminal 1935 essay, “The Work of Art in the Age of Mechanical Reproduction.” They followed this thread, encountering questions that mirror our current fears and curiosities about AI in science fiction novels and films over the years. Students explored the cyborg imaginary in Frank Herbert’s The Eyes of Heisenberg (1966), sentient decision-making in Arthur C. Clarke’s 2001: A Space Odyssey (1968), and technological dystopia in James Bridle’s New Dark Age (2018).

Benjamin famously found the “aura” present in painting—an authenticity unique to original works of art, which cannot be replicated in reproductions—to be absent from early photography. Mohaiemen’s students, taken with Benjamin’s idea of the “aura,” did what any good pupil would: they tried to prove him wrong. Utilizing the camera, photoshop, and a range of AI software like Midjourney and Stable Diffusion, students attempted to produce images with their own “aura.” This resulted in several projects that interrogated students’ own personal archives.

One such student is Leah Solomon, an MFA candidate with focus on sculpture and an interest in furniture-making. She was inspired by letters she found addressed to her uncle during the Eritrean War for Independence in the 1970s that revealed his background in furniture making and carpentry. Curious about the sort of furniture he would’ve made at that time, Solomon fed her own organic designs, along with found images, into Midjourney to create images that she would use as the basis of a sculptural project. The end result was a table with an embedded 3D design which melded elements of Eritrean furniture design (based on Midjourney’s speculations) with her own design aesthetic. Her own family archive being scarce, AI presented an opportunity to cobble together her own semi-fictional heritage.  

“AI often gets the critique that it’s sort of creating a history or manipulating images where there’s already history,” said Solomon, “but my particular interests in it were to generate these images where there otherwise would have been none, and in that way it’s speculative… I don’t think, in the context that I was using it, that it was harmful or disparaging to the history that I was working with.” 

Similarly, Photography MFA student Francisco Javier Ramirez used AI to further his goal of building new queer utopias in a post-AIDS world. “Because we have not been given the opportunity to create our own narratives and to be present in the present,” he said, “AI has these capabilities of allowing us to create a different sort of archive.” For him, this archive took the form of collage, bringing together real and fictional archival images. In 1990, Ramirez’s father emigrated from Mexico City to Utah and converted to Mormonism. Ramirez used AI to contrast the clichés of the American missionary experience (the software generated portraits of conventionally attractive young white men with picturesque mountains in the background) with his and his chosen family’s realities, as queer people and people of color.   

In Mohaiemen’s pedagogy, students drive the discourse. Supplementing Google’s grant, the Provost’s grant allowed Mohaiemen and his students to invite leading figures in the field to give public lectures in a series they called Machine Visions. Speakers faced an audience of discerning, critical art students who did not shy away from sticky issues concerning the public discourse around art and AI—such as the ethics of “scraping” visual artists’ work for machine learning training sets and the use of automation in roles traditionally held by people. “Considering the challenging questions our students asked about automation and its dark side (the possibility of mass unemployment and resulting global unrest), some technology startups do not ultimately have a satisfactory answer about the ‘why.’” Mohaiemen said. 

The class dove into these issues, which too often go unconfronted by tech start-up leaders, discussing which creative industries are at the highest risk of mass unemployment at the hands of AI. They landed on fantasy novel illustration, since this type of image does not necessarily need to make serialized, sequenced sense—still a weakness of AI photographs, Mohaiemen explained—and AI engines can quickly and easily pull from the sum total of past fantasy book covers to create a “new” image.

“Imagine the illustrator who was handed volume one of JRR Tolkien’s Lord of The Rings to read in 1955, with the instruction to dream up Mordor and Sauron,” said Mohaiemen. “The ultimate act of invention out of whole cloth, even though Tolkien drew from familiar mythology elements.” If you look at artist Alan Lee’s illustrations for the series, it is hard to deny that they have an aura. But if AI can generate an aura, or something close to it, does it matter that it is not the work of one individual artist? A class action lawsuit against Midjourney filed by a group of artists argues that it does. In a new age of mechanical reproduction that would likely cause Benjamin to implode, how do we value the individual imaginations of artists? Can AI be a part of this—and if so—can it be done ethically? 

Complex questions like these inspired Mohaiemen to pivot the focus of his Machine Visions lecture series from hosting AI industry leaders to hosting artists experimenting with AI in their daily practice, such as Trevor Paglen, Alexey Yurenev, and CAMP. About this transition, Mohaiemen noted, “We found the artists, both invited speakers and our students, to be more inventively curious and, simultaneously, healthily skeptical. They are not attached to any imperative to always be ‘eternal sunshine’ about this space.”

These parallels are evident in the work of the students and speakers, who often employ AI to create archives for marginalized communities who have historically been left out of, or removed from, the capital-A-Archive. Alexey Yurenev, for instance, used machine learning softwares to create a fictional archive of World War II photographs. Having grown up in Moscow within the echoes of the war, he felt that the uncanny nature of the synthetic images struck at something potentially more honest about war, and his own grandfather’s experience of it, than the official Soviet portraits (which, through elaborate staging and costuming, are often evasive in their own way). Like Solomon and Ramirez, Yurenev considered AI as a tool he could use to think deeply into a complex problem—rather than as a quick-fix solution.

Often, for artists, this means using AI as a way to disrupt, rather than generate meaning. Akira Kawahata, another MFA student in AI and Photography, entered the Photography concentration wanting to move away from certain strictly formalist traditions in Japanese photography. In his project for the class, Kawahata combined hyper-detailed photographs he had taken of vegetables with AI generated visuals of human skin that took on his vegetable surface. The resulting images, extremely sharp in visual quality, were obscure in their content. He created quite lively vegetables; a once-smooth cabbage leaf featured bumpy pores and intricate, risen veins. 

For Kawahata, AI didn’t serve to agitate the archive, but to reshape his work on the conceptual level. “I invite some element[s] generated by AI, based on my prompt, and it's deeply connected to concept—it's kind of collage photography,” he said about the project. 

Much like the camera once was, AI is a new tool for art-making, and it doesn’t seem to be going away. And as demonstrated by Mohaiemen’s students and speakers, it can be a useful and exciting tool. But perhaps it is most useful when considered and utilized alongside existing approaches—less so when it is valorized as the best new thing or denounced as the worst—as Solomon put it, when “it becomes a part of an artist’s practice, but not necessarily the most obvious part.”

So, what are Mohaiemen’s hopes for the future of AI in art and in the world at large?

“‘Humanity first’ needs to be our mantra here,” he concluded, “against the hyper-market credo that technologies that can generate mass unemployment do not need any regulation and oversight.” 

Mohaiemen and Ramirez will be participating in the “AI and Creativity” panel, alongside boundary-pushing Columbia colleagues across schools and disciplines, at the Columbia AI Summit on March 4, 2025.