A new research paper proposes a community-based auditing approach to identify and address AI harms specifically affecting queer communities. The study highlights how standard technical audits, which often rely on aggregate measures, can overlook identity-specific harms like erasure and discrimination. By involving queer community members in a participatory workshop, the researchers were able to classify harms across the AI lifecycle and trace them to issues in problem definition and data collection, as well as problematic contractual structures. AI
IMPACT This research highlights the need for more inclusive AI auditing methods to prevent discrimination against marginalized groups.
RANK_REASON Academic paper detailing a new methodology for AI safety research. [lever_c_demoted from research: ic=1 ai=1.0]
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
