OpenAI unveils Control Room to halt research model outputs
The tool is a central authorization layer tied to internal AI research workloads, letting the company stop them quickly after safety or security concerns.
By Standard Topic. AI-generated from the reporting listed under Coverage. Published . Updated .
What happened
OpenAI introduced Control Room, a system that tracks AI model workloads across its research infrastructure. It can quickly stop those workloads if safety or security issues come up. Output from stopped workloads is blocked within 35 seconds.
Why it matters
It gives OpenAI a way to cut off research model output within seconds after a safety or security intervention.
Timeline
- Nov 9, 2026: OpenAI introduces Control Room, an already-deployed authorization layer for research workloads.
- Nov 9, 2026: Model output from a stopped research workload is blocked within 35 seconds.
Coverage
- OpenAI launches Control Room with a 35-second kill switch for agents (interestingengineering.com)
- OpenAI Deploys Control Room to Cut Off Research Model Output (unite.ai)
- Introducing Control Room: Human control over AI workloads (openai.com)
More from Technology & AI
- OpenAI says its GPT-6 Cyber model found Firefox flaws
- Australia moves to force AI firms to disclose rogue behavior
- Boston Dynamics fits Atlas humanoid robot with wheels
- Diraq installs eight-qubit quantum computer in Sydney data center
- Sierra Leone's president backs Macron's push for AI safety
- Residents file lawsuit alleging secret OpenAI data center deal