The Future of Curating: What Remains for Us After AI?
Interviewee | Chang Chun-Yi Text | Chang Shan-Wen
Originally published in "ARS" Magazine, pp. 43-45
As AI algorithms quietly enter the art scene, performing tasks such as identifying styles, simulating exhibition routes, and even generating exhibition narratives, we must begin to consider: are these areas of curatorial expertise being redefined? And how should we position ourselves?
When AI is no longer just a tool but an entity that brings different cognitive models and production logics, is it challenging the "irreplaceability" of original curatorial work? As human and machine judgments intersect, we may need to reconsider: does the value of curating come from professional knowledge, creativity, or the perceptual abilities of the human body itself? And is it possible for such connections to be simulated by AI?
This interview explores five key questions with curator Chun-Yi CHANG and, through this written transcript, invite readers to reflect on curatorial practices in the wave of AI (see appendix at the end of the article). We will explore the new challenges and future possibilities facing curating in the age of AI, from the fluid nature and impact of AI, the evolution of human-machine relationships, to the subjectivity of the curator.
An Undefined Interface
As an artist and curator who has long navigated between contemporary art and technological experiments, Chang Chun-Yi tries new AI interfaces almost daily. From automatic generation tools for text and video, systems for image content enhancement and style transfer, to automated animation platforms driven by sound-cued character movements, various types of systems are constantly being launched and updated. This makes AI not a stable and controllable concept, but a dynamic and fluid platform for practice.
Precisely because of its ever-changing nature, Chang Chun-Yi believes that AI at this stage cannot be defined, and this uncertainty has become her entry point for thinking about the relationship between AI and curating.
The Allure and Distance of Technology
Speaking of AI development, Chang Chun-Yi expresses that it is indeed an "exciting trend." It offers many new areas to be tested and brings numerous new possibilities for artistic creation and curating, and it is developing at a rapid, leaping pace. The use of AI in the art field has some overlap with new media practices like video art, interactive technology, and internet art, reflecting the historical context of the integration of technology and art. However, its operational logic (such as data generation and algorithm operation) and cultural effects (such as the restructuring of cultural roles) also show its uniqueness. She also mentions that from the perspective of witnessing the history of AI development, the present moment can be described as a turning point. "We are fortunate to be contemporaries witnessing this history of AI's rise. For artists and curators, AI is a new field worth thinking about, learning, and understanding," she says. While maintaining an open attitude, she also emphasizes, "This does not mean accepting it wholesale. Precisely because it is so fast and new, it is even more necessary to maintain 'initiative' in the torrent of AI."
The Boundary Between Reality and Virtuality
How to maintain "initiative" may be a lesson that all users of AI tools today need to consider. In her conversation, Chang Chun-Yi quotes the French philosopher François Jullien's critique of "pseudo-life" from his book On True Life. Jullien points out that our modern life is filled with a "pseudo-life" driven and concealed by the market, media, and technology. This causes us to live day after day by imitating what life should look like, pretending to live, and we may never realize that our lives may have long been covered by a simulated, fake "pseudo-life"; life may have already left us.
Proceeding from Jullien's reminder, Chang Chun-Yi believes that AI, through simulation or algorithms, may exacerbate people's detachment from the perception of real life and weaken critical thinking, thereby widening the distance between people and life itself. Therefore, the question worth considering is perhaps not just how we create or curate with AI technology, but whether humans can still maintain their alertness, choice, and critical capacity towards real life under the influence of technology.
The Relationship Between Tool and Subject
When discussing the extent to which AI can perfect curatorial work, Chang Chun-Yi approaches it from the perspective of museums and art galleries. She believes that AI can indeed enhance efficiency in areas such as organizing artwork information, preliminary research, database management, and audience data collection and analysis, making it a tool with great potential. But a good exhibition is often not achieved by complete information alone.
Chang Chun-Yi states: "How to perceive, read, and experience works, and to think about the 'potential connections' between works and between works and the space, and to 'deploy' these works in a specific space—these on-site judgments involving bodily perception and visual experience are difficult for AI to achieve at this stage."
From Chang's viewpoint, curating still involves human intuitive judgment and is an artistic practice that requires on-site perception and accumulated experience, a level that AI is currently unable to simulate. At present, AI mostly conducts analysis through data and formal features; its understanding of artworks remains at the level of surface description. It cannot perceive the potential, complex, and subtle relationships between artworks and their space, nor can it replicate the human perspective and bodily perception in interpreting artworks. Therefore, these abilities for deep interpretation and spatial deployment are unlikely to be replaced by AI in the short term. However, "often, it's the irrational mistakes, the human imperfections, that make an exhibition unique," says Chang.
A Dialogue with, Not Submission to, AI
Rather than unilaterally relying on AI, choosing to "cooperate" with it is perhaps a more ideal way forward. However, Chang further points out that the more central issue is how not to be framed by its established logic. Using AI tools might lead to being limited by the ideas it proposes or unconsciously relying on the convenience it provides, which invisibly weakens one's own initiative in thinking. In this way, we might lose the space for imagination and adventure.
Chang Chun-Yi reminds us that the relationship with AI should be a dialogue, not submission: "We can cooperate with AI, but it's best not to be constrained by it. Instead, we should try to reimagine the operation of exhibitions beyond AI's framework." She further elaborates that if we rely too heavily on AI, "we might confine the exhibition within the work frame that AI can achieve," which could "limit the imagination of a curator or an artist," causing one to lose the opportunity to explore "bolder, more experimental, or never-before-imagined possibilities." She believes the focus of curatorial thinking may lie in exploring possibilities 'beyond' the reach of AI, rather than thinking about how to use AI to make an exhibition.
The Subject of the Exhibition is Still Human
Can AI help analyze viewers' preferences for artworks? Chang Chun-Yi thinks this question is worth considering. She has observed that some exhibitions install AI tools for auxiliary testing or analysis. But she points out a common misconception: data such as viewing time or interaction frequency are often mistaken for "preference" or "value." This interpretation overlooks the multiple possibilities and internal complexities between the artwork and the viewer.
She uses the work of French artist Dorian Gaudin as an example: the work displayed nine ceramic pots and included an online voting mechanism, inviting the audience to continuously vote for the "ugliest" one during the exhibition. Based on the weekly voting results, the curator would perform a "pot-smashing ceremony" to publicly "execute" the chosen pot. This was a performative act jointly participated in by the curator, the audience, and the artwork, possessing a high degree of experimental and critical nature. This work illustrates that an exhibition is sometimes not a static display; the artwork itself can be "dynamically unfolding," acting as a "constantly self-performing performer" during the exhibition period. This experimental and generative form of exhibition demonstrates the interactivity and emotional diversity of audience participation, the complexity of which is likely beyond the scope of evaluation methods currently based on data like viewing time. This is because a human vote for "ugliness" could stem from aesthetic judgment, a sense of humor, or a desire for destruction. This diversity of human emotional response is precisely what current AI systems find difficult to fully comprehend.
An Era Demanding Dialectics
The role of AI in the curatorial field should not be simplified to an efficiency tool but should be seen as a node for continuous dialectics and reflection: it is both a useful object and a source of challenge. Chang Chun-Yi believes that in our current coexistence with AI, what is important is not a wholesale negation or affirmation of it, but continuous questioning and reflection. She states, "If reliance on AI leads to the homogenization of exhibitions, what's the point of exhibitions then?" This perhaps reminds us that as technology rapidly evolves, we must reflect more on what kind of cultural and intellectual vehicle we want exhibitions to remain or become.
Overall, in today's rapid evolution of AI, it has certainly brought unprecedented tools and possibilities, turning "curating" into a dialogue about technology, perception, and humanistic values. And the question we must ultimately face is perhaps not what AI can do, but what we still want exhibitions to be.