OpenAI 推出 Control Room，35 秒内可叫停模型工作负载
OpenAI 把所有训练评估用的推理任务接进了一个统一管控系统，发现异常 35 秒内就能停掉任务。做大规模智能体采样训练的团队可以看看它的管控思路，和 Anthropic 的安全栈相比这更偏运行时干预。
OpenAI 发布了内部系统 Control Room，用于追踪研究基础设施上的所有模型工作负载，并在出现安全或安全隐患时快速叫停任务。系统从干预决策到停止或暂停模型输出的时间在 35 秒以内。
Control Room 接入了每一项研究推理工作负载，包括用于模型训练和评估的智能体采样、数据处理等环节。它与 OpenAI 的模型对齐异常监测系统协同工作，一旦监测系统判定存在风险，Control Room 即可对相应任务执行停止或暂停操作。
该系统是 OpenAI 前沿模型安全训练与部署整体方案的一部分。随着前沿模型的训练推理规模扩大，这类集中式工作负载管控机制正在成为头部实验室安全基础设施的组成要素。
We built Control Room, a central system to track model workloads across our research infrastructure and stop them quickly when safety or security concerns arise. It connects every research inference workload, including agentic sampling for model training and evaluation, data processing, and more. This works together with our misalignment monitoring systems, and halts model output within 35 seconds of a decision to intervene to stop or pause a job. Control Room is just one part of our holistic approach for safe frontier model training and deployment. openai.com/index/control-…📊 25 ⚡
This report is categorized under AITOP's "Products" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison and AI regulation, compliance frameworks, and sovereign compute policy. Key entities and technical topics under continuous tracking include OpenAI, Control Room, AI安全, 智能体.
Source attribution: curated from OpenAI (@OpenAI). Cross-source corroboration: 11 independent sources have covered this development (including theverge, marktechpost, Patrick Loeber), confirming industry-wide signal strength.
