“被上千种分类法害死？”：实践中的AI风险分类
"Death by a thousand taxonomies?": AI Risk Classification In Practice
浏览论文内容
中文总结 AI 辅助
本文通过对25位多领域人士的访谈，开展SOT开发与使用的实证研究，发现SOT在AI治理中整合度低，分析其原因并提出设计建议，指出需完善治理基础设施以发挥SOT潜力。
中文摘要 AI 辅助
AI所涉及的危害在性质和范围上各不相同，从不安全的用户交互到AI应用带来的全社会层面的后果。对AI的各类风险进行分类是AI治理的基础：监管机构、科技公司和政策制定者需要结构化的风险描述来作为行动依据。因此，研究人员和从业者开发了许多社会技术结果分类法（SOT）。本文对SOT的开发和使用进行了实证研究，研究基于对来自工业界、学术界、民间社会和政府的研究人员和从业者开展的25次访谈。我们发现SOT在AI治理流程中的整合度较低，并确定了SOT设计和使用的两个特征来解释这一现象。首先，SOT用于生成AI风险这一复杂问题空间结构化表示的设计选择，往往对下游分类法用户不可见，这些用户将生成的类别视为风险的详尽描述，而非解释工具。其次，SOT通常仅列举危害，未将其与决策点或涉及危害发生的行动者关联起来，导致难以分配问责。最后，我们为SOT开发者和用户提出了设计建议，并指出要实现SOT的潜力，需要目前尚未存在的治理基础设施。
英文摘要
The harms in which AI is implicated range in nature and scope from unsafe user interactions through to the societal-wide consequences of AI adoption. Classification of the diverse risks of AI is foundational to AI governance: regulators, technology firms, and policymakers need structured accounts of risk upon which to act. Researchers and practitioners have accordingly developed many Sociotechnical Outcome Taxonomies (SOT). This paper presents an empirical study of SOT development and use, drawing on 25 interviews with researchers and practitioners across industry, academia, civil society, and government. We find SOT are weakly integrated into AI governance processes, and identify two features of SOT design and use that explain why. First, the design choices through which SOT produce structured representations of the complex problem space of AI risks tend to be invisible to downstream taxonomy users. Those users treat the resulting categories as exhaustive accounts of risk rather than as interpretive aids. Second, SOT typically enumerate harms without linking them to decision points or actors implicated in their occurrence, leaving accountability difficult to assign. We close with design recommendations for SOT developers and users, and argue realising the potential of SOT requires governance infrastructure that does not yet exist.
