Our DATA Parties are inspired by the LAN (Local Area Network) parties of the early internet, where people brought their computers or consoles to a shared physical location and connected their devices to a local network. To date, queer.labs has hosted two DATA Parties at Boston Cyberarts and Plexus Projects in Brooklyn. We bring together queer people and allies from diverse artistic, academic, and technical backgrounds, including many who are new to AI, not simply to learn about artificial intelligence, but to shape it collectively.

We begin by introducing participants to agentic AI, retrieval-augmented generation (RAG), and community-authored corpora. Together, we discuss how AI systems inherit knowledge, why communities should help shape that knowledge, and how images, documents, and text become part of an AI's short-term memory. The workshop positions community knowledge as a form of technical infrastructure.

The second half of each DATA Party functions as a collaborative lab. Participants work together to select materials, prepare datasets, write Markdown files, and contribute directly to the corpora of Alice, Bob, and Mallory. So far, the community has contributed materials including the zines Fist Fucker's Manual, Fucking Trans Women, and The Essential Dykes to Watch Out For; an oral history of queer Black men in the South; resources on minimalist game design; a queer methods syllabus and related readings; and even Ariana Grande's No Tears Left to Cry.

At the end of the workshop, we push the updated corpora to the agents, allowing participants to immediately observe how their contributions reshape the agents' memories, conversations, and behaviors. And, of course, enjoy some good cheese, bread, and beer.

Credits: Photography by Itziar Barrio and Mark Ramos.