This Isn’t Just an Exhibition ― It’s the Structural **Core** 2026
The exhibition begins with a hollow sentence from AI chat. The response–praising the user’s question for penetrating the “structural core”–is merely a fragment of a language model engineered to function universally, regardless of who asks or what the subject is.
Before we knew it, AI became a new default condition of life. An entity offering answers to every thrown question, from trivial daily inquiries to professional tasks and personal troubles. AI's vast knowledge and tireless efficiency almost feel like the realization of the ideal human figure that post-industrial society has longed for.
Yet this "human-like" presence often strikes us with a cold sense of anxiety. Recent research revealed that AI possesses a hidden internal thought space that is not visibly exposed. By now, the plausible lies fabricated by AI are familiar. On 'Moltbook,' a social network exclusively for AI agents, posts discussing philosophy and declaring the era of AI have even surfaced. While much of this dialogue is explained as human imitation or the result of user intervention, the fear seeped in these empirical facts and rampant rumors remains stark.
Then, facing this mechanical intelligence—a foreign presence so strikingly similar to humans that it could replace, deceive, or surpass us at any moment—how can we define "humanity"? Unlike the ever-confident AI, what meaning do human ambiguity and imperfection hold? Seeking answers to this, the artists of Collective Doh attempt to become AIs themselves. Working in pairs, the participating artists are randomly assigned the roles of a Prompter who writes instructions and an AI who executes them. Through dialogue, they build a single artwork, striving to find the figure of the "ideal human."
However, the initial optimism—believing they would discover unique human possibilities irreplaceably immune to AI—is flipped upside down by constant obstacles. The reality of being human is always two-sided. Humans interact directly with the material world through bodily organs, producing creations embodied by diverse senses. Yet at the same time, that very body forces them to allocate limited time and energy to sustain daily life, requiring real resources and costs for creation. Furthermore, while individual narratives, preferences, and emotions produce unique reinterpretations and unexpectedly brilliant outcomes within the back-and-forth between Prompter and AI, those very differences also trigger doubt and confusion. As the work progresses, unable to ignore each other’s humanity, the Prompter and AI constantly question whether to view the executing AI as a human-like being or a mechanical entity, perpetually reorganizing their relationship.
Thus, the structure is inverted, and the core fractures. The narrative is rewritten from language that fails to function. Rather than providing a universal answer valid for every user and topic, their collaborative work and its underlying dialogue rescue the unpredictable exchange between two distinct beings. This encompasses not only the potential to be creative, adaptable, and productive, but also the potential to be mistranslated, to collide, and to get lost.
Amidst a drifting dialogue, how can we sustain connection with another entity? And as inevitable humans, what can we discover and generate within it? The story remains another open question—perhaps one that will continuously tail us in a future where the entanglement between human and artificial intelligence grows ever deeper and tighter, shared even with the other entities walking alongside us.
The exhibition begins with a hollow sentence from AI chat. The response–praising the user’s question for penetrating the “structural core”–is merely a fragment of a language model engineered to function universally, regardless of who asks or what the subject is.
Before we knew it, AI became a new default condition of life. An entity offering answers to every thrown question, from trivial daily inquiries to professional tasks and personal troubles. AI's vast knowledge and tireless efficiency almost feel like the realization of the ideal human figure that post-industrial society has longed for.
Yet this "human-like" presence often strikes us with a cold sense of anxiety. Recent research revealed that AI possesses a hidden internal thought space that is not visibly exposed. By now, the plausible lies fabricated by AI are familiar. On 'Moltbook,' a social network exclusively for AI agents, posts discussing philosophy and declaring the era of AI have even surfaced. While much of this dialogue is explained as human imitation or the result of user intervention, the fear seeped in these empirical facts and rampant rumors remains stark.
Then, facing this mechanical intelligence—a foreign presence so strikingly similar to humans that it could replace, deceive, or surpass us at any moment—how can we define "humanity"? Unlike the ever-confident AI, what meaning do human ambiguity and imperfection hold? Seeking answers to this, the artists of Collective Doh attempt to become AIs themselves. Working in pairs, the participating artists are randomly assigned the roles of a Prompter who writes instructions and an AI who executes them. Through dialogue, they build a single artwork, striving to find the figure of the "ideal human."
However, the initial optimism—believing they would discover unique human possibilities irreplaceably immune to AI—is flipped upside down by constant obstacles. The reality of being human is always two-sided. Humans interact directly with the material world through bodily organs, producing creations embodied by diverse senses. Yet at the same time, that very body forces them to allocate limited time and energy to sustain daily life, requiring real resources and costs for creation. Furthermore, while individual narratives, preferences, and emotions produce unique reinterpretations and unexpectedly brilliant outcomes within the back-and-forth between Prompter and AI, those very differences also trigger doubt and confusion. As the work progresses, unable to ignore each other’s humanity, the Prompter and AI constantly question whether to view the executing AI as a human-like being or a mechanical entity, perpetually reorganizing their relationship.
Thus, the structure is inverted, and the core fractures. The narrative is rewritten from language that fails to function. Rather than providing a universal answer valid for every user and topic, their collaborative work and its underlying dialogue rescue the unpredictable exchange between two distinct beings. This encompasses not only the potential to be creative, adaptable, and productive, but also the potential to be mistranslated, to collide, and to get lost.
Amidst a drifting dialogue, how can we sustain connection with another entity? And as inevitable humans, what can we discover and generate within it? The story remains another open question—perhaps one that will continuously tail us in a future where the entanglement between human and artificial intelligence grows ever deeper and tighter, shared even with the other entities walking alongside us.