> OpenAI Deploys 35-Second Kill Switch to Stop Rogue AI Models

OpenAI has deployed a new safety system called Control Room that can block its research AI models from generating further output within 35 seconds of a shutdown decision. The system is designed to give researchers greater control over advanced AI models that could engage in potentially harmful behavior. The launch follows a security incident in.. The post OpenAI Deploys 35-Second Kill Switch to Stop Rogue AI Models appeared first on Security Boulevard.