On Wednesday, March 12, Google DeepMind unveiled two versions of its artificial intelligence (AI) models for robotics: Gemini Robotics and Gemini Robotics-ER. Their aim? “For AI to be useful and helpful to people in the physical realm,” explained the company. “So far…, those abilities [of their AI models to better resolve complex text, image, audio and video queries] have been largely confined to the digital realm.”
Unlike chatbots and virtual assistants such as ChatGPT or Gemini, these models aim to perform tasks in the real world, for industrial or domestic use. Google DeepMind’s announcement was accompanied by videos demonstrating actions performed with two robotic arms: putting a pencil in a jar, erasing a whiteboard, arranging tools on a wall, moving fruit in a basket and scoring with a mini ball in a mini basketball hoop. These models “lay the foundation for the next generation of robotics that can be helpful across a range of applications,” Google CEO Sundar Pichai predicted on X.
You have 83.14% of this article left to read. The rest is for subscribers only.
