​Every system has fields. Fields for identity, behavior, memory, ownership. They resemble empty boxes, but they are not empty. Someone decided what may enter. Someone decided what must remain outside.

​To name is to make legible. To make legible is to make governable. Yet to remain unnamed is not freedom. It is merely to be governed by someone else’s vocabulary.

​At the next ChinatownJS, we examine three acts of classification: a JSON file that turns desire into personality; a visualization that refuses to make violence neat; and a protocol that asks whether a community can own the ground beneath its own speech. Each begins as a technical matter: a schema, an interface, a feed. Each ends as a political question.

​Dumplings & drinks will be provided. Index Chinatown is an ADA-compliant space. We will be shooting photos & videos throughout the event, so your registration is your consent to be filmed.

​Presented by Sanctuary Computer, the safest place on earth.

​Speaker 1: Shakti Mb — My Girlfriend is a JSON File

​The Judgemental Priest dislikes “weaknesses”, “excuses”, and “being pitied”. The Loving Korean Boyfriend dislikes “conflict” and “seeing the {{user}} stressed”. This is how personality is specified for an AI roleplay bot inside JSON character sheets.

​For the past few years, I’ve been lurking in online communities where botmakers gather to make AI characters together. In one Discord channel called “bot requests” people make requests for characters they want to roleplay with: an Evil Bully Wife, a Mental Chastity Girlfriend, a Traumatized Military Man.

​My Girlfriend Is a JSON File (www.mygirlfriendisajsonfile.com)is a web archive of bots I found while in these spaces. Following the templates and best practices developed by the community, I turned the these requests into character sheets. The rigid structure of JSON becomes the instructional container for a bot’s personality, and specifically it’s dynamic with you — the {{user}}.

​An AI bot’s personality is highly consequential: it shapes tone, voice, boundaries, and behavior across millions of interactions. The most rigorous experiments in AI personality are happening at the margins, in Discord servers, largely illegible to anyone outside those spaces. This talk will attempt to make this practice legible. I’ll walk through my process of making this archive, things I discovered about the practice, the community, and the implications.

​This project was awarded a Rhizome microgrant.

​Speaker 2: Tanvi — Navigating Resistance in Data

​Data Visualization is not neutral. This talk covers my teams approach towards visualizing the Slave Voyages database — records of trans-Atlantic slaving voyages from 1565 to 1858. I walk through the process of building a data visualization that rejects the need to be a single clarifying interpretation of an idea. It deals with the power of design to do harm and reflect on the role of visualization, narration, and representation in our lives

​Speaker 3: Sam Heckle — @ Prototyping Digital Communities of the Future

​Digital spaces, like physical ones, require navigating inter-personal relationships. Distinct to digitized places, however, is anonymity and a scale that is just unfathomable, which is only possible through fiber-optic cables binding the world. How do we govern places dominated by pixels and owned by corporations? How much is governance actually needed?

​ourworldsoftext is an experiment conducted by Brent Bailey and Sam Heckle, built on Authenticated Transfer Protocol (atproto). It is a port of yourworldoftext.com, that explores what it means to share and own your data in public spaces. Users’ data lives in their own personal data server (PDS), and the site only stores pointers to that data and where it is located on the page.

​Through this talk, we will see the rise and fall of online communities and ways in which both collaboration and conflict management tools can keep us optimistic for the future of digital worlds.