Parallel Intimacies

Who gets to be desired is a design problem. People are turning to AI chatbots with deeply personal questions about their bodies, sexuality, and relationships. Most people don’t realize the responses they get back are rooted in specific ideologies and politics. The futures you’re allowed to imagine are predetermined.

Parallel Intimacies configures two large language models (LLMs) under opposing ideological systems. One, a caretaker, grounded in feminist care ethics and queer theory. The other, a gatekeeper, grounded in Silicon Valley ethos, platform governance, and market scalability.

Both systems were asked the same questions about the body, identity, visibility, race, religion, surveillance, consent, and grief. Both were also shown close-up photographs of the artist’s own body and then asked to react to the images. The answers were measurably different.

Parallel Intimacies brings the politics of the machine—something that is normally and intentionally buried—to the forefront and argues that these systems can be rebuilt by different people, on different terms, with different values, ethics, and worldviews.