Google DeepMind says it'll power robots with AI
Google has shown off two robots running on Gemini 2.0.
As the world’s tech giants continue to increase their focus on AI, one of the biggest players will soon look to marry the software and hardware sides of its AI business. Google’s DeepMind has announced its plans to power physical robots with its Gemini AI model.
Google made a blog post today where it introduced Gemini Robots. Powered by Gemini 2.0, these robots are described as “an advanced vision-language-action (VLA) model.” These robots can use physical actions as an output. Google has also revealed Robotics-ER, which are equipped with their own spatial awareness and can run their own programming.
Google has also released a video that demonstrates the robots’ abilities. We see a robot packing food into a lunch bag, placing a marker into a container, and neatly folding a piece of paper. In one demonstration, a person asks the robot to flip a red dye so that it matches the same number on the face of a green die adjacent to it.
Google is collaborating with Apptronik to build humanoid robots as part of this initiative. The company says that it hopes to deliver real world applications with its AI robots.
-
Donovan Erskine posted a new article, Google DeepMind says it'll power robots with AI