AI Studies Reading Group

Crawford's text is accessible and easy to enter, and maybe because of that, the conversation ran well over an hour. After the first session's more foundational, theoretical reading, this one pulled us into the power dynamics of AI that drew most of the group here.

The session began with a summary of Crawford’s key arguments and its relavent ideas and references. But what I wanted to share is artificial intimacy. This week, one of the articles stuck with me for a while. So I pivoted the theme a bit toward artificial intimacy and care within the datafication of human beings as resources through Hyodol, the AI "care dolls" for lonely elderly people in South Korea. Someone noted the doll is never only a companion — it monitors vitals, detects when someone has died alone, routes medical data to social workers; care folded together with management. We sat with the uncomfortable parts: that being cared for carries its own vulnerability and indignity, and that it might be easier to be in those states with a machine you owe nothing to. This has a long prehistory — the role of the mother delegated to media, from the baby monitor onward (cf. Mother Media) and an MIT scholar Sherry Turkle has tracked artificial intimacy since her early experience in a nursing home and a new book, Artificial Intimacy (upcoming, 2026).

What is intimacy between a human and a machine, and how does it change over the centuries? Delegating the role of something too vulnerable, the ick(!) moments — think of a doll speaking "I love you" from a device — becoming an acceptable substitute? Why do humans delegate such roles to machines? What does it mean when this gets extreme, and the feeling of intersubjectivity in the physical world becomes too rare, and difficult to reach?

Then emotion. One of the striking references from a participant is a challenge circulating on Instagram and TikTok right now, where people say the same sentence in four or five emotions and everyone joins in because it's funny.

We ended on the blurring of truth and objectivity — the sense that the "scientific," objective ground we stand on is itself built on something fictional, on world-building and storytelling — which is exactly where the next session goes. Our presenter left us with Borges' beautiful quotation.

...In that Empire, the Art of Cartography attained such Perfection that the map of a single Province occupied the entirety of a City, and the map of the Empire, the entirety of a Province. In time, those Unconscionable Maps no longer satisfied, and the Cartographers Guilds struck a Map of the Empire whose size was that of the Empire, and which coincided point for point with it. The following Generations, who were not so fond of the Study of Cartography as their Forebears had been, saw that that vast Map was Useless, and not without some Pitilessness was it, that they delivered it up to the Inclemencies of Sun and Winters. In the Deserts of the West, still today, there are Tattered Ruins of that Map, inhabited by Animals and Beggars; in all the Land there is no other Relic of the Disciplines of Geography. — Jorge Luis Borges, Collected Fictions, trans. Andrew Hurley

The group kicked off with an energetic vibe with artists, film critics, translators, script and game writers, and digital-rights researchers across Seoul, Berlin, New York, Manila, India, and beyond.

We opened with Toister and Zylinska's essay. After the introduction, I realized that most of us are interested in its geopolitical power dynamics, but I wanted to begin at a more foundational level: how can we understand AI? My own recurring curiosity is how AI is reshaping human epistemology; how we see, how we know, and, most of all, how we feel the world in this shifted era. Most importantly, art-making under such conditions. Zylinska's earlier text, one of our upcoming readings, takes up exactly this epistemological turn; this co-authored essay sharpens it by anchoring the question in the relationship between text and image, which is close to my own research.

Because AI is built on the mechanism of the neural network, it can be situated within a long genealogy in which technology reflects the functions of the human organ, and we, as humans, realize the exact mechanisms of the organ afterward, retroactively. The question that follows, for me as an art historian, is what becomes of epistemology — so often centered in the humanities — once it is run through this loop. One of the main skills that my discipline, art history, requires is to describe artworks and objects through text; I keep returning to ekphrasis, the ancient Greek rhetorical practice of conjuring an image through words and language. If the essay is right that generative AI performs a kind of "cognitive hacking," what does that do to ekphrasis, to the act of putting the visual into language? It's a question I have been revisiting from my own work as a curator, and sometimes as an art critic for a long time.

"if the architecture of human thinking remains opaque, how can its artificial counterpart be engineered with confidence? How can thinking be made visible, including to the thinker, and what part do images now play in this process?" — Toister & Zylinska, 432

On Hara Shin’s Work and Abstraction

During the talks, I asked: do we all think that we sense this changed, peculiar visuality? If so, is it just because of AI-generated or mediated works, or something else? Like a cognitively or epistemologically informed visuality? Hara Shin, an artist based in Berlin, shared Spiral Slime and Oblique Ellipse (2021), a GAN-based VR work that unexpectedly produced an "abstracted landscape." She had set two datasets against each other — one close to the "real" (clear weather, nature, the body), the other its fabricated counterpart. Training drove the error close to zero, yet the image went off track and nearly collapsed: instead of the photoreal landscape she expected, she got something gestural and almost painterly. Abstraction has long been understood as form liberated from representation — think about the post-war abstract paintings — and human painters arrive there by deliberate choice. On the other hand, Shin's GAN arrives at the same place by failing to represent.

What does it mean to reach the same destination by opposite paths, one by will and the other by collapse? How even is it possible? Is something else at stake? I haven't yet pinned down the logic of that failure, which is, I think, exactly why it's worth staying with.