Google DeepMind’s most recently upgraded AI model has bestowed robots with the capability to undertake more intricate tasks and even glean information from the internet. Carolina Parada, who leads DeepMind’s robotics division, revealed at a press conference that the company’s novel AI models operate in a collaborative manner. This collaboration allows robots to engage in “multi-step reasoning” prior to taking real-world actions and to meticulously plan their courses of action in advance.
Backed by Gemini Robotics 1.5 and the embodied reasoning model Gemini Robotics-ER 1.5, this system represents an updated iteration of the AI model introduced in March. Presently, robots are not limited to performing simple, singular tasks like folding paper and opening bags. They are now capable of sorting clothes based on dark and light hues, packing suitcases in accordance with weather conditions, and aiding people in categorizing waste, compost, and recyclables, all guided by the results of online searches.