Google's DeepMind says it will use AI models to power physical robots

Core Insights - Google is integrating its DeepMind AI technology into robotics, showcasing its capabilities at the annual I/O developers conference [1] - The company introduced two new AI models, Gemini Robotics and Gemini Robotics-ER, which are designed to control robots and perform physical actions [2] AI Models and Capabilities - Gemini Robotics operates on Gemini 2.0, which Google claims is its "most capable" AI to date, extending beyond text and images to physical action commands [2] - The AI models for robotics must be general, interactive, and dexterous to be effective in real-world applications [5] Partnerships and Collaborations - Google is partnering with Apptronik, a Texas-based robotics developer, to create next-generation humanoid robots using Gemini 2.0 [3] - Apptronik has previously collaborated with Nvidia and NASA, and Google participated in its recent $350 million funding round [3] Demonstrations and Applications - Demonstration videos showed Apptronik robots executing tasks like plugging in devices and filling lunchboxes in response to spoken commands [4] - The timeline for market release of this technology has not been disclosed [4] Industry Context - Google is not alone in the pursuit of AI for robotics; OpenAI has invested in Physical Intelligence to develop general-purpose AI for robots [6] - Other companies, including Tesla, are also entering the humanoid robotics space, indicating a competitive landscape [7]