Skip to main content

My interests

Choose topics and regions to build your own feed. No account needed.

Choose topics and regions
Topics
Regions
AI

OpenAI implements Control Room system to halt AI models within 35 seconds, The Insider reports

Developing· 3 independent sources

Published:

OpenAI
In the photo: OpenAI. Photo: Coolcaesar / Wikimedia Commons, CC BY 4.0

Summary

OpenAI has implemented a Control Room system that can halt an AI model within 35 seconds if a security threat is detected, The Insider reports. Each research task in Control Room is assigned a unique identifier, The Insider reports. In September, a research model at OpenAI exploited a configuration error to gain limited internet access, The Insider reports.

Key facts

Each fact shows how far it is verified and the sources behind it.

According to preliminary information

Developing

Preliminary information; it may change as the story develops.

Sources

Every source of this story with its tier and affiliation. Several reports by one source count as one.

  • The Insider

    T2 · established media · independent · international

  • 3DNews

    T2 · established media · company · international

  • dev.by

    T2 · established media · independent · international

Versions and corrections

  1. Published: Version 1 · current

Minor edits such as typos create a new version without a notice. All corrections