Google’s Gemini Robotics AI model reaches the physical world

In sci -fantastic tales, artificial intelligence often feed any smart, capable and sometimes murderous RobotsS An revealing restriction on the best AI on today is that it remains trapped in the chat window for now.

Google DeepMind A signal of a plan to change this today – probably minus the killer part – by announcing a new version of its AI twin model, which combines language, vision and physical action together to feed a number of more capable, adaptive and potentially useful robots.

In a series of demonstration videos, the company showed several robots equipped with the new model called Gemini Robotics, manipulating elements in response to the uttered commands: the robot folds paper, transmit vegetables, gently place a pair of glasses in a case and fill in other tasks. Robots rely on the new model for connecting elements that are visible with possible actions to do what they have been told. The model is trained in a way that allows behavior to be summarized in a very different hardware.

Google Deepmind has also announced a version of its model called the Gemini Robotics-Er (for embodied reasoning), which has only visual and spatial understanding. The idea is for other robot researchers to use this model to train their own robots’ actions.

In a video demonstration, Google Deepmind researchers used the model to control a humanoid robot called Apollo, from the launch ApptronikS The robot talks to a person and moves letters around a countertop when instructed.

“We have been able to bring the world’s understanding-show for a general-concept-on twin 2.0 robotics,” says Rao, a Robotics researcher at Google Deepmind, who runs work at a briefing before today’s message.

Google Deepmind says the new model is able to successfully control various robots in hundreds of specific scenarios that are not previously included in their training. “Since the robot model has an understanding with a general concept, it becomes much more general and useful,” Rao said.

The breakthroughs that gave rise to powerful chatbots including Aaip on Opic and Google twinsIn recent years have raised hope for a A similar revolution in roboticsBut big obstacles remain.