Unsupervised Pleasures - Intersectional Language Models for Queer Futures

Explore the intersection of large language models and queer representation in this interactive workshop. Delve into the datasets shaping AI systems like GPT-4, examining how they represent diverse communities and identities. Analyze existing datasets from an artistic research perspective and discover collaborative efforts to build alternative models centered on decolonial perspectives, queer love, ethics of care, and mutual aid practices. Engage in discussions about developing more intersectional machine learning methodologies, addressing issues with current tools, and envisioning inclusive AI systems that support underrepresented communities. Suitable for all levels of technical expertise, from beginners to experts, this 54-minute session encourages participants to imagine and contribute to the creation of language tools that truly reflect the diversity of human experiences.