What kind of expression would the knowledge we humans receive be wearing?
AI Trainer Kim’s Life | Solo Exhibition
November 2025
by Konno Yuki
Before leaving for work, I eat breakfast while still half-asleep. I hate wasting time wondering what to eat, so the menu is fairly monotonous—three kinds of bread, two kinds of jam, and a salad. I bring my cup back from the desk where I was working late last night. I pour some milk into a cup of coffee. I mumble once, still not fully awake, wondering what the bread would say just before it gets eaten by the butter. I tell myself that my health depends on my choices, and at the same time I want to escape this state where I am forced to choose between this and that. These days I feel that way even on days that are not particularly busy. The age of fixed TV channels has passed; now I just watch whatever the algorithm floats up. As I read the news, I see so many stories about young adults; that must be because of the gender and age I entered. The appearance of keywords like art and labor is probably because of what I searched a few days ago. I am not actually interested in illness, archaeology, or relics, but those are clearly the traces of what came up while I was gathering materials the other day. Before I even start to think, my finger taps a link and I read the article without thinking. We live in an era where empathy is selective and compassion is as plentiful as view counts, and yet I... I mutter things like that to myself as I close the webpage. A little while ago, somewhere, I picked up the line that modern people should have room for imagination instead of blind choices. It was not something I chose to listen to, but I find myself wanting to respond to it: No, imagination and choice have already become one, and here imagination can only be a simulation grounded in reality. So, what do you think? Today, once again, I wear myself out as I move back and forth between imagining situations that could exist and ones that might never exist.
Areumbit Park’s solo exhibition ≪AI Trainer Kim’s Life≫(2025) begins with the specialized profession of the AI trainer. AI trainers are the people who register the information that is used as answers to users’ questions in AI assistant services. Though they remain unseen, the AI we have come to regard as our tutor of knowledge operates as a kind of composite knowledge system. As the questions we ask grow more diverse and specialized expertise is required, AI trainers are called in. In the text that shares its title with the exhibition, we can follow not only the conversations between AI and its human users, but also the thoughts on the everyday life and labor of AI trainers, presented together with images. Drawing on interviews and the artist’s own experience, the text asks: what kind of existence are AI and the AI trainer? And more broadly, how are contemporary forms of labor and knowing constituted through technology?
In this exhibition, Areumbit Park addresses how conversations with AI assistants inevitably touch upon moral and ethical dimensions. How are morals and ethics, which set the standards for a shared life, formed through the answers provided by AI? Even a single question can produce different answers, because it is impossible to fully separate the responses entered and generated by trainers from their own sociopolitical perspectives. In the video work ≪What Would Bread Say Right Before Being Eaten by Butter?≫(2025), the large question Who is God? is followed by a range of responses. Some reject the existence of God, some insist on a single, absolute deity, others claim that God resides in natural objects, and still others indicate that who or what counts as God can differ depending on context. These answers may offend some and satisfy others. Here, answers based on the individual trainer’s personal beliefs are not allowed; instead, AI is expected to provide customized responses that fit the user’s own views and assertions. Through such conversations with AI, humans mobilize knowledge that appears both universally valid and finely stratified, and in this process of learning, users come to accept these responses as what counts as knowing.
What kind of expression would the knowledge we humans receive be wearing?
AI Trainer Kim’s Life | Solo Exhibition
November 2025
by Konno Yuki
Before leaving for work, I eat breakfast while still half-asleep. I hate wasting time wondering what to eat, so the menu is fairly monotonous—three kinds of bread, two kinds of jam, and a salad. I bring my cup back from the desk where I was working late last night. I pour some milk into a cup of coffee. I mumble once, still not fully awake, wondering what the bread would say just before it gets eaten by the butter. I tell myself that my health depends on my choices, and at the same time I want to escape this state where I am forced to choose between this and that. These days I feel that way even on days that are not particularly busy. The age of fixed TV channels has passed; now I just watch whatever the algorithm floats up. As I read the news, I see so many stories about young adults; that must be because of the gender and age I entered. The appearance of keywords like art and labor is probably because of what I searched a few days ago. I am not actually interested in illness, archaeology, or relics, but those are clearly the traces of what came up while I was gathering materials the other day. Before I even start to think, my finger taps a link and I read the article without thinking. We live in an era where empathy is selective and compassion is as plentiful as view counts, and yet I... I mutter things like that to myself as I close the webpage. A little while ago, somewhere, I picked up the line that modern people should have room for imagination instead of blind choices. It was not something I chose to listen to, but I find myself wanting to respond to it: No, imagination and choice have already become one, and here imagination can only be a simulation grounded in reality. So, what do you think? Today, once again, I wear myself out as I move back and forth between imagining situations that could exist and ones that might never exist.
Areumbit Park’s solo exhibition ≪AI Trainer Kim’s Life≫(2025) begins with the specialized profession of the AI trainer. AI trainers are the people who register the information that is used as answers to users’ questions in AI assistant services. Though they remain unseen, the AI we have come to regard as our tutor of knowledge operates as a kind of composite knowledge system. As the questions we ask grow more diverse and specialized expertise is required, AI trainers are called in. In the text that shares its title with the exhibition, we can follow not only the conversations between AI and its human users, but also the thoughts on the everyday life and labor of AI trainers, presented together with images. Drawing on interviews and the artist’s own experience, the text asks: what kind of existence are AI and the AI trainer? And more broadly, how are contemporary forms of labor and knowing constituted through technology?
In this exhibition, Areumbit Park addresses how conversations with AI assistants inevitably touch upon moral and ethical dimensions. How are morals and ethics, which set the standards for a shared life, formed through the answers provided by AI? Even a single question can produce different answers, because it is impossible to fully separate the responses entered and generated by trainers from their own sociopolitical perspectives. In the video work ≪What Would Bread Say Right Before Being Eaten by Butter?≫(2025), the large question Who is God? is followed by a range of responses. Some reject the existence of God, some insist on a single, absolute deity, others claim that God resides in natural objects, and still others indicate that who or what counts as God can differ depending on context. These answers may offend some and satisfy others. Here, answers based on the individual trainer’s personal beliefs are not allowed; instead, AI is expected to provide customized responses that fit the user’s own views and assertions. Through such conversations with AI, humans mobilize knowledge that appears both universally valid and finely stratified, and in this process of learning, users come to accept these responses as what counts as knowing.