基于社区的酷儿群体AI危害审计
原标题：Queering the Audits: Community-Based Auditing of AI Harms to Queer Communities
AI 摘要
研究者在 EurIPS 2025 的 Queer in AI 工作坊组织了一场参与式审计活动，16 名酷儿社区成员使用 4Cs 危害分类法（Content、Conduct、Contact、Contract）审查了心理健康聊天机器人、广告支持聊天机器人、内容审核和数据标注劳动四个案例。参与者将危害追溯到问题定义与数据收集环节，并识别出从弱势群体中提取价值却几乎不提供救济的合同结构。研究指出，社区参与的审计能揭示聚合评估方法容易忽略的身份特定危害。
正文节选
Queering the Audits: Community-Based Auditing of AI Harms to Queer Communities Abstract AI systems embed majority-group defaults into training data, evaluation metrics, and category definitions, producing documented harms for queer communities including erasure, misclassification, and discrimination. Standard technical audits often rely on aggregate measures and cannot detect harms that become visible only through the lived experience of affected communities. We conducted a participatory auditin