Long before AI imitated artists, Vera Molnár became a human computer

As a student on the Hungarian University of Fine Arts, Vera Molnár learned all the standard tricks of easel painting: symmetry, balance, pyramidal composition. She was not impressed. Even many years after graduating in 1947, Molnár couldn’t consider that “such a prestigious national institution could turn out to be so pathetic and mediocre.” But for all her griping, aesthetic conservatism provided her with almost all the pieces she needed to turn out to be probably the most radical and original painters of her time.

Molnár’s extraordinary achievement is now honored in a posthumous Retrospective on the Centre Pompidou in ParisInstalled just months after her death on December 7, 2023 – shortly before her one centesimal birthday – the exhibition is each a historical homage and a challenge to contemporary art.

Like many avant-garde artists of her time, Molnár defined her work as a “countermovement.” Her statement was deliberately broad. The act of opposition was initially more necessary than what she was opposing. Armed with a store of artistic ideas that she had collected on the Hungarian University of Fine Arts, Molnár was well prepared to wreak havoc. What made her unique was her systematic approach. From 1960 onwards, she set about deconstructing the “cultural readymades” of art history using a pc.

The problem was access.

Shortly after graduating from art school, Molnár had an inkling that a pc program could process visual information in ways in which a classically trained artist would never have imagined. The algorithmic rigor of a machine would produce formal arrangements unconstrained by historical antecedents, let alone the aesthetic assumptions that artists aspired to within the name of culture. Unlike their flesh-and-blood counterparts, computers weren’t graduates of the Hungarian University of Fine Arts, the Art Students’ League, or the Royal Academy.

However, computers were also prohibitively expensive. Universities and governments were lucky in the event that they had a single mainframe computer. Scientists and engineers competed fiercely for precious minutes of computing time.

In contrast, access to programming languages ​​was essentially free. Anyone could buy a book with instructions on how you can program in BASIC or FORTRAN. Molnár learned these languages ​​and wrote easy programs that she ran in her head and recorded the outcomes on paper. For a couple of years, Molnár became a pc. She called her invention a imaginary machinean “imaginary machine”.

Almost a decade later, Molnár was given access to an IBM System/370 and a Benson drum plotter on the Sorbonne Computing Center in Orsay. At first she worked unofficially on evenings and weekends, benefiting from the proven fact that the French research institute was caught up within the protests of 1968. She operated a real machine was exertions, probably harder than the imaginary machine which she was well accustomed to. Mainframe computers didn’t yet have monitors, so she needed to submit a stack of punch cards and hope for fulfillment a couple of days later.

Naturally Success was not easily defined of their system. Conventional aesthetic appeal would have been a failure, but Molnár was not searching for the Dadaesque randomness of a Hans Arp collage. The underlying motivation, uniquely original, was to agnostically explore a variety of possibilities after which resolve subjectively which ones had unexpected coherence.

Arp didn’t put his name on every uncontrolled arrangement of scraps of paper. The difference is that he curated the arbitrary, while Molnár deliberately selected from an exhaustive set of possibilities. To reinforce her selections, she also often added one other layer of interpretation by repeating her drawn lines as drawings or paintings.

“Thanks to these combinatorial possibilities,” wrote Molnár, “the computer enables systematic research in the visual field.” The key to this statement is the word Researchwhich she pursued through her artistic practice (often supported by her husband, an artist turned computer scientist). Her artistic research was necessary from each the visual arts and computer science perspectives.

The artistic significance is obvious in hundreds of paintings and drawings, but is especially evident in her engagement with the Bauhaus master Paul Klee. In 1969, when Molnár gained control of IBM and three many years after Klee’s death, the Musée d’Arte Moderne in Paris held a serious Klee retrospective. One of the paintings Molnár saw was called Variations (1927). She took the title as a challenge and spent the subsequent few years writing variations Variations in her head and on the mainframe. In a way, Klee was resurrected. (“Paul Klee owns me,” she explained.) On a deeper level, Molnár pioneered a way that was simply impossible in Klee’s lifetime, an actual gap in what art could express.

Molnár was at the very least as necessary as a researcher in computer science. Remarkably, her importance is way greater today than it was when she made her contribution: Molnár’s art created a relationship with technology that was largely forgotten or missed by Silicon Valley tech bros, especially in the sector of artificial intelligence.

The extraordinary pattern recognition of generative AI and the provision of computers exponentially more powerful than the IBM System/370 have given AI the looks of a cultural behemoth. From video to literature to music, the corpus is way larger than any human could experience in a lifetime.

Molnár lived long enough to understand this power, but she rarely used it. Her approach was based on responsibility, which began along with her experience within the black box of the mainframe computer and involved careful consideration of initial conditions and decisions.

This allowed Molnár to effectively control the system on a holistic level. The feedback loop ran through her brain and infrequently through her hands. Any reinforcement of prejudice was linked to accountability; inattention rationalized and reified the production of cultural readymades.

Has the technological dynamic of Microsoft, Google and Meta turn out to be inevitable? The confirmation bias inherent in generative AI, which incessantly promotes itself, suggests that we’re doomed to be cogs in a planetary real machine. However, art researchers within the tradition of Molnar are well positioned to intervene by reviving the imagination and reopening the feedback loop that Molnar pioneered in 1968.

Molnar often repeated an aphorism by Paul Klee that “art is a flaw in the system.” The flaw shouldn’t be an error, but slightly a break that reveals hidden errors and offers space for correction.

Instead of entering the black box of Dall-E or Chat GPT and accepting their unknowable unknowability, art researchers should attempt to turn out to be AIs imaginaryIn Molnár’s absence, the position is open to anyone who’s willing to take temporary possession of it.