Gemini Robotics 2 Brings Google's AI Into the Physical World
Wired
Read full postGoogle DeepMind launched Gemini Robotics 2, an AI system that integrates vision and action models to control robots performing complex tasks like screwing in lightbulbs and tidying shelves. The system combines vision language and action models to interpret surroundings and execute physical movements, marking a step toward physical artificial general intelligence (AGI). While promising, the deployment of such AI-controlled robots raises safety concerns due to potential unpredictable behaviors.


