Tortured in Logic and Irrelevant
Chances are you first encountered generative artificial intelligence as I did, with the release of ChatGPT in November 2022. The tech podcasters at the time were abuzz about it, causing me some peripheral awareness of its arrival. My awareness mushroomed when several students in my Fall 2022 Humanities course submitted take-home exams in December featuring responses that were tortured in logic and irrelevant to our work. The responses bore no similarity to previous writing they had produced in our class. I immediately suspected that ChatGPT had been used.
Unfortunately, our quick turnaround from the fall to the spring semester, alongside new initiatives we were implementing at the start of the spring, didn’t allow me to investigate my suspicions. Instead, the students received the grades that ChatGPT likely deserved in its handling of existential, essential Humanities questions: a “D.”
I queried our Academic Integrity working group at my university about my experience. Our campus had started a foray into generative artificial intelligence detectors, which led us to discover the false positives the detectors were generating, and thus our subsequent abandonment of the detectors. The university then convened a ChatGPT Working Group comprised of mostly faculty alongside a few administrators. We were charged to investigate how other institutions and professional disciplines were addressing the use of generative artificial intelligence and provide our campus faculty with recommendations for how to attend to the presence of this entity within academics. Given my background in the arts, I focused on the generative artificial intelligence’s impact on the arts. I was also driven to investigate the ethical considerations in using generative artificial intelligence in both the classroom and the workplace. For this, I sought guidance in the form of coursework.
Taking Pause
I enrolled in an online course with the Node Center in Berlin, Germany, devoted to Artificial Intelligence and Art and taught by Marco Mancuso, a critic, curator, researcher, and professor at Fine Art Academy in Bergamo specializing in art, philosophy and technoscience. I had taken several courses at Node over the years and was intrigued by this offering as its description posed the question of “what ethical and critical questions arise from new human-machine relations?” (Node Center for Curatorial Studies, 2010-2020, Artificial Intelligence and Art with Marco Mancuso Description section, para 1).
One of the most substantive takeaways from this course for me was revisiting the primacy of human intelligence in how we encounter the world. This emerged as a key ethical consideration as it was the first time I had taken pause to consider the existence of other intelligences. I was introduced to the work of James Bridle, a writer, artist, and technologist who actively thinks about the multitude of intelligences including planetary intelligence. For Bridle, rethinking and reintegrating our relationships with the world removes the hierarchy of human intelligence within the world (Shavron & Bridle, 2022). This reimagining and reordering of intelligences helped me in removing some of the sense of fear, associating artificial intelligence with replacing humanity. The recognition of other intelligences including those of the animal, vegetal, fungal, bacterial, and planetary varieties, reduced my anxieties about the likelihood of robots overtaking the world.
Another takeaway from my time at the Node Center centered on the discovery of the profound aesthetic possibilities of working with generative artificial intelligence. While media art and digital art have used automated systems in creative practices for several decades, these mediums are now using machine learning and large datasets of information in creative production. The work of Refik Anadol, a new media artist and designer, offered me a substantive and thought-provoking aesthetic experience.
The Life of Buildings
Refik Anadol’s (2023) body of work “invites his audience to imagine alternative realities by redefining the functionalities of both interior and exterior architectural elements” (Biography section, para. 3). His institutional installations bring buildings to life. “WDCH Dreams” (2018) instilled the Walt Disney Concert Hall, home to the L.A. Philharmonic, with a consciousness that stirred within me a deepened sense of the life of buildings.
As a former actor who had worked in theater spaces across New York City, my fellow actors and I would often enter a historic performance space in awe. We often wondered aloud what the walls would say if they could speak. These were structures that had housed legendary performers in their time. Some of these awe-provoking spaces included the Duplex cabaret in the West Village in New York City that features the early performances of Joan Rivers and Barbra Streisand; the Opera House at Brooklyn Academy of Music where Isadora Duncan, Enrico Caruso, Sarah Bernhardt and Arturo Toscanini tread the boards; and the Victory Theater in Times Square which at one time was home to Lionel Barrymore and Lilian Gish in the long-running “Abie’s Irish Rose.”
Anadol’s “WDCH Dreams” (2018) was created for the centennial celebration of the L.A. Philharmonic that used both its Frank Gehry-designed hall and one hundred years of images, music, videos, and printed programs (nearly 45 terabytes of data) to remember and honor the Philharmonic’s accomplishments. The data were transformed using machine learning and projected onto the exterior of the concert hall during the opening week of the celebration. The resulting piece was a 12-minute “combinatorial fantasia” (Rose, 2018, para. 10) that can be viewed on YouTube.
Within the installation is a passage titled “Dreams,” which made me weep in front of my computer monitor. This was developed through Anadol’s engineering to cause the machine learning model to hallucinate. Not only had the walls of the concert hall been talking in the earlier passages using the historical data culled from the work of conductors Esa-Pekka Salonen, André Previn, Zubin Mehta, Gustavo Dudamel and others, but the walls were now dreaming of other realities. Perhaps the dreams were of these conductors as well as past musicians and performances. It was this element of imagination—a building that dreamed!– that touched on the intimate relationships I had with the performance spaces in which I worked.
The Human Experience
The experience of witnessing buildings possessing sentience might be a curious way to relieve some of my fears surrounding the proliferation of generative artificial intelligence in our lives. But it was this creative work between humans and machine learning that helped me understand aspects of my humanity. And it was the unalloyed aesthetic experience of that encounter, with my jaw hanging open, my breathing paused, and the tears rolling down my cheeks, that opened me to the positive contributions that can be made in using generative artificial intelligence. But the full experience requires human collaboration along with human viewers to make meaning of the effort.
Generative artificial intelligence is now a reality in our lives. One of the better strategies in harnessing it is to use it. During my studies with the Node Center, we created storyboards using both text and images generated by large language models. My storyboard used essential questions regarding social justice as prompts. In March 2023, these prompts proved to be confounding for DALL-E 2, resulting in error messages without results. By September 2023, the generator was producing images that were adequate, but not evocative of the human experience embedded in the questions.
My university continues to research the ways in which artificial intelligence can contribute to our academic practices. The recommendations that our campus’s Center for Teaching and Learning have offered include the use of project-based strategies in our curriculum that can be enhanced with the use of generative artificial intelligence but not wholly produced by it. Strategies like these require the use of authentic assessment, with students demonstrating the application of knowledge they gained in a course. It involves real-time, human-to-human encounters and exchanges.
Perhaps the greatest takeaway from my exploration of generative artificial intelligence is that its existence could heighten the need, the importance, and the desire for human interaction and reflection on the human experience.
Note: To view Refik Anadol’s work cited in this article, I recommend a 5-minute trailer, “WDCH DREAMS – Process” located at this link: https://vimeo.com/307533505. The full performance (15:02) can be viewed at this link: https://youtu.be/FKfCrChDWpY?si=rv3iw0jnLjXbvq21.
References
Anadol, R. (2023). Biography. Refikanadol.com. https://refikanadol.com/information/.
Shavron, S. and Bridle, J. (26 October 2022). James Bridle — planetary intelligence. Media Evolution. https://www.linkedin.com/pulse/everything-intelligent-recognising-diversity-intelligence-/.
Node Center for Curatorial Studies. (2010-2020). Artificial intelligence and art with Marco Mancuso. NODE Curatorial Studies Online. https://nodecenter.net/course/learn/jBBbinG6LPCJpgRx5.
Rose, F. (2018, September 14). “Frank Gehry’s Disney Hall is technodreaming.” The New York Times. https://www.nytimes.com/2018/09/14/arts/design/refik-anadol-la-philharmonic-disney-hall.html
