Google DeepMind Unveils Advanced AI Models to Enable Robots to Browse the Web for Task Assistance

Google DeepMind has unveiled advanced AI models, Gemini Robotics 1.5 and Gemini Robotics-ER 1.5, enhancing robots’ capabilities to tackle complex tasks and utilize online resources. During a recent briefing, Carolina Parada, head of robotics, revealed that these models enable robots to “think multiple steps ahead,” allowing them to perform diverse actions, such as sorting laundry by color, packing suitcases based on weather, and classifying waste through location-specific web searches. This marks a shift from executing single tasks to genuine understanding and problem-solving in physical environments. The upgraded Gemini Robotics-ER 1.5 interprets its surroundings and employs Google Search to synthesize information into understandable instructions for robots. Additionally, robots can now share skills across different models, leading to seamless functionality between various robotic configurations. Google DeepMind is rolling out Gemini Robotics-ER 1.5 via the Gemini API in Google AI Studio, with select partners accessing Gemini Robotics 1.5.

Source link