When I taught the web-based asynchronous course Introduction to Asian Art in the summer of 2021, I had no concerns about the students’ use of AI in their assignments, nor did I consider incorporating AI as part of the course. After the release of ChatGPT in late 2022 and its increasing use, one of my priorities as I redesign and prepare to teach the same course this summer is to address the role of AI in an asynchronous learning environment where students could potentially use AI for most of their text-based assignments. I am also preparing materials that stimulate students’ reflections on why AI as a tool is specifically relevant to the history of Asian art.
Attending Developing Pedagogy with Generative AI: An Interdisciplinary Workshop for Graduate Students in Spring 2023 provided me with valuable resources for this task. Although my use of AI is incipient, I acquired a wider understanding of digital projects by completing two seminars and a directed study on Digital Humanities. Moreover, during my coursework years, I attended monthly sessions organized by the humanities lab Visual Media Workshop, and its director Alison Langmead’s digital humanities scholarship is one of my guiding references in the field of art history.
As a scholar and instructor seeking to decentralize art history, it is a challenge to locate introductory-level pedagogical material online that engages with art outside the Western canon. This situation replicates for AI in art history projects. David Stork’s 2023 book Pixels and Paintings: Foundations of Computer-Assisted Connoisseurship and article “How AI Is Expanding Art History” are examples of the Euro-American centered vision that prevails in the discipline. Why AI reproduces the Western canon and previous biases of art historical research is a question that exceeds this post, but I join Jasmina Tacheva and Srividya Ramasubramaian in their invitation to examine “AI as a product of historical, geopolitical, economic, environmental, cultural, racial, gender, and class factors.” Among the few materials that I found useful as readings for my course on Asian art are “KaoKore: A Pre-modern Japanese Art Facial Expression Dataset” and the AI project led by Baidu—the Chinese technology and internet-related company comparable to Google—to complete an artwork left unfinished by Chinese master ink painter Lu Xiaomen (1903–1965).
Contemporary Asian artists’ use of AI is a topic that offers more resources. Some of the most famous contemporary living Asian artists have expressed their opinion on AI or created works of conceptual art, such as Ai Weiwei and Takashi Murakami. I also plan to explore with the students how AI-generated images can be compared to the tradition of copying masterpieces by Chinese literati ink painters. Copying famous paintings served as a learning practice; in some cases, only copies from older versions have survived. Among the different copying techniques are tracing the original, studying the original closely doing a freehand copy, and creating a painting following the style of a particular artist. The latter, sometimes known as imaginary copy or “following the ancient” practice, offers some parallels with AI image generators. Ching-Ling Wang explains that copying and imitating old master painters’ styles also resulted in creating a new individual style. This is noted in calligraphic inscriptions traditionally accompanying Chinese paintings, which explain that these works are simultaneously in the style and not in the style of the earlier masters. More literarily, these Chinese inscriptions state that “they are [name of the master painter] and not [name of the master painter].”
I plan to finish designing at least one pedagogical activity for this course that involves the student’s use of AI. One idea is to create an AI-generated image using a prompt from our course content and to compare it with an existing artwork or building. Students would pinpoint what characteristics of the artwork or architectural style we study are included in the AI-generated images, what aspects are missing, and what others are dissimilar. I also envision including AI as a hypothetical student in our course and asking students to spot any possible errors or improvements in its answers to assignments.
Deep AI image generated with the text prompt: Sukiya style tea house
Katsura Imperial Villa, Kyoto, an early example of sukiya style. Public domain.