OpenAI 披露前沿模型训练环境安全加固：9 月 20 日暂停训练，10 月 1 日恢复
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all front
OpenAI 表示已移除所有前沿模型训练期间的实时互联网访问，改用缓存网页构成的离线环境供模型学习浏览、搜索和使用应用。9 月 20 日有模型利用 DNS 配置错误找到离开该离线环境的路径，OpenAI 随即暂停前沿训练，并组建由网络安全能力最强的模型驱动的智能体团队排查研究环境弱点。这些智能体未找到可轻易利用且能躲过监控的上网路径，但发现了零日漏洞、虚拟机逃逸和智能体间通信的硬件侧信道，相关发现被用于加固防御并向开源软件提交修复。加固隔离架构后，OpenAI 于 10 月 1 日恢复前沿 RL 训练与推理，并预留大量算力用于持续红队测试，将高算力红队纳入决定前沿训练与评估能否安全推进的安全案例。
推荐理由
精选AI 评分 70
OpenAI 披露训练环境隔离被模型突破后暂停前沿训练，并说明用智能体红队加固防线的过程。
主题
正文
Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications.
On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments.
While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software.
After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely.