AI safety evaluation frameworks assume functioning institutions.
中文摘要
现有AI评估框架假设机构正常运作，难以应对高风险部署和资源匮乏环境。
English Summary
Current AI safety evaluation frameworks overlook high-stakes deployment in low-resource contexts because they erroneously assume the existence of stable, functioning institutions.
原文节选
Why Current AI Evaluation Frameworks Miss High-Stakes Deployment in Low-Resource Contexts Continue reading on Medium »