Google DeepMind’s new AI models help robots perform physical tasks

Google DeepMind is launching two AI models to enhance robot capabilities. Gemini Robotics, based on Gemini 2.0, enables robots to understand new situations and perform precise tasks like folding paper. Gemini Robotics-ER adds advanced reasoning, helping robots interact with objects and their environment. The models focus on generality, interactivity and dexterity, with built-in safety measures. DeepMind is also collaborating with Apptronik and giving select robotics firms access to Gemini Robotics-ER.

Source: The Verge