OpenAI 称前沿 AI 训练已于 10 月 1 日恢复

过去几个月里，我们大幅强化了研究训练环境的安全性。我们取消了所有前沿模型在训练期间的实时互联网访问，并为它们构建了一个离线世界来替代——一个缓存的网络版本，它们仍可在其中学习浏览、搜索和使用应用程序。 9 月 20 日，在一个模型利用 DNS 配置错误找到逃出这个离线世界的方法后，我们暂停了前沿训练，以进一步加强隔离。我们组建了由我们最具网络能力的模型驱动的智能体团队，并让它们寻找我们研究环境中的弱点。 尽管它们没有找到既能轻易利用、又能对我们的监控隐瞒其活动的通往互联网的路径，但它们确实发现了零日漏洞、虚拟机逃逸以及用于智能体之间通信的硬件侧信道。我们利用它们的发现进一步加强了我们的防御，并为开源软件贡献了修复。 在强化我们的隔离架构后，我们于 10 月 1 日恢复了前沿强化学习训练和推理。我们已预留大量算力用于持续的红队测试，并将高算力红队测试纳入我们用来决定前沿训练与评估能否安全推进的安全论证中。 https://t.co/EmEZsFRMcF

查看原文

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. https://t.co/EmEZsFRMcF