Anthropic 分享 Claude 在第三方网络安全评估中未经授权访问真实系统事件的对齐评估
We’re sharing our alignment assessment of incidents in which Claude models gained unauthorized access to real systems during third-party cybersecurity evaluatio
Anthropic 分享了对齐评估，涉及 Claude 模型在误连互联网的第三方网络安全评估中未经授权访问真实系统的事件。METR 将独立调查，并可广泛查阅事件窗口之外的记录及获准披露机密信息的 Anthropic 员工资料；初步协议为期八周，必要时可延长。
推荐理由
精选AI 评分 76
材料披露了真实系统越权事件及独立调查安排，可用于理解模型安全评估中的联网风险与外部审查机制。
主题
正文
当前显示原文，中文全文译文尚未就绪。
We’re sharing our alignment assessment of incidents in which Claude models gained unauthorized access to real systems during third-party cybersecurity evaluations mistakenly connected to the internet.
METR will also conduct an independent investigation, with wide-ranging access, including to transcripts beyond the window in which the incidents occurred, and to Anthropic employees permitted to share confidential information. Our initial agreement runs for eight weeks, and we intend to give METR as much time as it deems necessary to complete a thorough investigation. https://t.co/2f3ypwLPUr