The Brooklyn Rail

For a while, I have avoided the word crisis in reference to our ecological condition because it seemed to provoke an intellectual standstill, the kind of dismay that concludes effort or even further consideration. Recent chatter about the AI crisis did the same. Likewise, the crisis in arts funding. There are even claims of a crisis in any social imaginary. Recently, however, I have begun to wonder if the permanent crisis of the humanities might not provide a different attitude towards crisis across these different areas.

The beleaguered humanities are notoriously ever in crisis. Taking a tip from the artist Carla Gannis’s regular queries on Google about various issues, searching “crisis of the humanities” delivered over 53 million results in 0.06 seconds. Books with that title date back sixty years. And yet, the humanities as a term does not even appear in the first edition of the Oxford English Dictionary in 1933; the concept is popularized after World War II and goes into its first crisis barely thirty years later. Various reasons are given for this permanent crisis, but one productive outcome has been to make its major thinkers and contributors doubt, debate, and defend its value using that rhetorical trinity of pathos, ethos, and logos. As they write, they also historicize the debate, introduce new perspectives and social conditions, compare disciplinary constructs, and even resist standard framing of the problem.

Easy claims to uplifting humanity are sentimental dross (even if sometimes true) and don’t work for defending the humanities in any meaningful way. I was once told as I was reading the history of philosophy and mathematics: “I don’t know why you want to study something that nobody cares about anymore.” I was in college and was suddenly made to realize that I would likely spend a lifetime explaining my intellectual choices. That hardly seems a bad thing. I have in that way constantly examined and revised how my words produce a lens on what I see, that is, present a landscape, or make a world.

A current show at the Ford Foundation, curated by Mashinka Firunts Hakopian and Meldia Yesayan, What Models Make Worlds: Critical Imaginaries of AI (on view through December 9, 2023), aims to disrupt general impressions about the machine learning algorithms of technology. A wall near the entrance presents the provocations of Caroline Sinders’s Feminist Data Set (2017–23) to help center intersectional, queer, Indigenous, and trans perspectives within a feminist framework because data collection, labeling and training occur within cultural and ideological contexts that are “a reflection of culture, of knowledge, of expertise, and data to come.” The fourteen artists (individual or collectives) in the show tackle the world making of AI systems.

The crisis of AI is most obvious in the technology’s ongoing bias concerning Black and brown people, a fundamental problem in continuing use of these major data sets; as one example, the Bureau of Labor Statistics finds 70 percent of fast food workers to be white but Stable Diffusion’s model represents fast food workers 70 percent of the time as Black or brown. Voicing Erasure (2020), a three minute video by Algorithmic Justice League, as well as works by Stephanie Dinkins, the recipient of the first LG Guggenheim Award earlier this year, or Astria Suparak’s Virtually Asian (2021) engage the stereotypes that seep into data sets. Researchers have for years critiqued the “dubious curation” and continue to emphasize the need for solidarity and ethical AI standards within these trillion parameter machine learning models; when people talk about the decreasing costs of training AI models as enabling greater accessibility, that depends on continued use of these problematic foundational sets like LAION. Scholars from leading research institutions proposed “The Foundation Model Transparency Index (FMTI)” necessitating greater transparency if AI companies are to be held accountable.

A work by the Iranian artist Morehshin Allahyari remains a stand-out in this conversation for the way she uses AI to model a better practice, reversing the gender bias introduced into Persian society by Western influence for her series “ماه طلعت، (Moon-faced)” (2022–23). In ancient Persian literature, “moon-faced” was a genderless adjective used to define beauty, whereas in contemporary Iran it refers solely to the beauty of women. Portrait paintings in the Qajar dynasty were historically characterized by a cross-gendered ideal of youthful beauty, but modernization, plus the increasing artistic and political exchange between the West and Iran, prompted the court’s embrace of realistic painting and photography, ending the prevalence of “gender-undifferentiation.” Using a multimodal AI trained on a Qajar dynasty painting archive (1786–1925), Allahyari reproduced the once-genderless portraits, resisting the Western gaze that was imposed upon the painting tradition.

Next to the video are two still works on velvet in frames with a peaked arch associated with Persian architecture. Velvet painting in the United States has associations with kitsch but painting on velvet has a global history and may have been invented by Islamic people, arriving in Europe during the Crusades, according to the scholar and curator Jennifer Heath in Black Velvet: The Art We Love to Hate (1994). The softness here provides the blurring sensibility of the shifting faces in Allahyari's AI generated video.

The “Moon-faced” series comes from an artist who is not uncritical of the many biases of AI generators but helps force recognition that the machine learning algorithms are ambiguous; they can also reintroduce aesthetics suppressed by a unilateral Western gaze. This complicates easy rejection and demands a more subtle attention to managing and regulating these technologies. Congress held the second AI Insight Forum on October 24, 2023 and a week earlier the Senate proposed the Artificial Intelligence Advancement Act of 2023 to address the technology’s use in finance services and the military. Increasing governmental attention doesn’t turn to its proliferation in our daily lives and so demands its citizen population become familiar with the issues.

Personalized AI assistants to support the merciless productivity demands of contemporary capitalism are expected to be widespread within the next couple years (and ChatGPT already offers approximations), so Lauren Lee McCarthy’s performance as a voice assistant in LAUREN (2017–ongoing) invites considerations of the class dynamics in such tools. Through the use of assorted surveillance systems, McCarthy unobtrusively observes and facilitates participants’ desires around their homes. Domestic workers for the upper class have always been “invisible” in this way, anticipating needs while also tasked to maintain cultural hegemonies for those they serve.

The virtual voice assistants are known to depend on underpaid global labor and reports earlier this year on the exploited labor of Kenyan workers scrubbing toxicity from ChatGPT also revealed how these AI systems would replace this underclass at the cost of those workers’ mental health and futures. The labor crisis is not ameliorated by these technologies, unless labor is designed differently. What Models Make Worlds is an important exhibition for audiences to view and engage with the social and economic issues that are pressing our representatives to make decisions; if we aren’t informed about these AI systems then how are we to voice our concerns and demands within this representative democracy?

“You can continue on your journey like a dachshund disguised as a poodle, but don’t forget to look closely.…” So says A’Yan in Lap-See Lam’s Dreamers’ Quay on view at the Buffalo AKG, in a humorous lesson about how mediation can alter expectations but responsibility remains with us to consider what we think about what we see. The shadow play projections immerse audiences in a film of a young girl named A’Yan moving across centuries through a dream world that abounds with cultural stereotypes and mediated images of China. The Swedish artist compiled 3D scans of her family’s Chinese restaurant and others as they were sold or closed, which became the basis for the fictional set, digitally rendered in sepia tones with characters as silhouettes.

This amalgamation of techniques and visual codes allows the misrepresentations of Chinese culture to become more obvious and inherent to mediation. The crisis of deep fakes is suddenly positioned within a longer history of representation. The black box at the renovated museum offers an ideal immersive experience with a curtained hallway that allows audiences to manage the transition both visually and intellectually. The work is particularly expressive within the context of the museum’s outstanding Photo-Secession collection; the movement, led by and associated with Alfred Stieglitz, rejected the documentarian attitude surrounding photography and emphasized the photographer’s manipulation for a subjective vision instead. It’s astonishing how frequently we still need to remember that with both photography and the growing realm of post-photographic practices.

At Borusan Contemporary Istanbul, Mat Collishaw also used an older technology, a zoetrope that works admirably for Sounding Sirens (2023), a six legged machine with an octopus in the center surrounded by jellyfish. As the lights flicker and the machine rises and falls, the octopus seems to be waving its tentacles at the onslaught of jellyfish, a reality of the climate crisis experienced within the aquatic ecosystem of the Bosphorus, where the work is on display. Appearances suggest that the surface dwelling jellyfish are overcrowding the solitary octopus, but that stems from the spinning machine and lights, thus inculcating human activity into this complex.

Collishaw adopts Google Colab’s Disco Diffusion to develop his “Dürer/Haeckel Image Generation” works (2023), merging the famous engraving of The Four Horsemen of the Apocalypse (1498) with Haeckel’s nineteenth century botanical and animal illustrations. The anxiety of Early Modern Europeans who witnessed Plague, War, Famine, and Death as warnings of end times contributed to the immediate success of Dürer’s woodcut. The artist is also credited with describing the basic principle of ray tracing (a technique used in computer graphics), which produces highly realistic visual renderings simulating an array of optical effects, so is nicely introduced into a wider conversation on the hyperrealism of image generators.

Collishaw adopts complex personalities but that strengthens a message of the show, which titled Arrhythmia speaks to the disjointed attitude of “civilization” to “nature” (as if not a part of it) but also the need for assessing the rhymes, if not the repetitions, of our histories. Dürer must be understood within a Renaissance interest in ideal human ratios; his Four Books on Human Proportion (1528) doesn’t attribute physiognomies to race or ethnicity and yet his work contributed to “the way visual strategies were used to parcel up humanity into distinct groups of operationally interchangeable individuals.” Haeckel’s illustrations inspired the Art Nouveau movement, but he also interpreted Darwin in now disproved racist frameworks, with some palpably anti-semitic positions, though his views on many matters led to his rejection by the Nazis.

In Collishaw’s “Dürer/Haeckel Image Generation” an ecological crisis appears, but so does the profound confusion we face in our intellectual and aesthetic histories. Our cultural predecessors were intermittently great and flawed so we can’t take their work simply, but must always contextualize its meaning for our own times. Dürer’s drawings show great interest in how figures might transform into one another and we see a popular fascination with the same in so much large model image generation. Collishaw notes that “AI itself has no form of its own, it takes on whatever shape it’s given and can easily run, or in this case writhe, amok in exaggerated objectivity.” The problem, as always, is our childish dream of a final and dependable take; we are instead obligated to keep questioning how, why, when, and from whom we adopt certain references or visual codes, logical and optical technologies.

Thinking in this historical vein, I was delighted to learn that Barry Threw and Gray Area has digitized the POINT Foundation’s Whole Earth Catalog and its many supplements from 1970 to 2002, made public as of October. Despite the various issues raised by scholars about how many of its founders and participants in this cultural project wound up enabling the fantasies of Silicon Valley, reading across the issues and supplements presents efforts to educate and limit extractive practices. Across the Whole Earth Index, authors inquire and investigate more often than they declare and determine. The perspectives inspire and exemplify where to do better. The Spring 1978 special publication, Soft-Tech, offered a surprisingly apt comparison of attitudes that still apply today. The message is one of thinking in terms of small not large, diverse not uniform, embracing resilience and adaptation.

Dürer wrote that “The ‘measure’ of a human figure is especially hard to comprehend, amongst other reasons because the human figure is composed neither by rule nor compass but is contained within irregular curved outlines, it is specially hard to write and treat of it.” Our crisis comes from a desire for certainty about ourselves, each other, and the world around us, but if we can relinquish that then suddenly opportunities present themselves. If we can accept that later everything we did will appear inadequate, if not an utter failure, then we can proceed from the assumption of necessitating different models and practices from time to time.

Uncertainty is uncomfortable but its destabilizing also allows us to question and reinvestigate, to hypothesize and experiment, to theorize and mobilize new ideas. That something has been, does not mean it should still be. That something was relevant does not mean it can’t be retired. That something was accepted does not mean it can’t be altered. The arts and humanities show us this time and time again, ever in crisis and ever surviving, and therein their thriving contemporaneity.