Carolina Parada, Google DeepMind's vice president and head of robotics, said embodied intelligence is the next frontier of artificial intelligence [1].

This shift represents a move from AI that processes digital information to systems that can physically interact with the world. By integrating advanced reasoning with physical movement, the company aims to create robots capable of performing complex, real-world tasks that previously required human precision.

Speaking on the sidelines of the Tokyo Humanoids Summit on May 29, 2026, Parada said the current direction of the Gemini Robotics platform [1]. She highlighted the development of highly dexterous robot tasks, citing examples such as packing a lunch box and folding origami [1]. These tasks require a level of fine motor control and spatial awareness that traditional robotics have struggled to achieve.

To support these goals, Google DeepMind has unveiled two advanced AI models for robots: Gemini Robotics and Gemini Robotics-ER [3]. These models are designed to provide the cognitive framework necessary for robots to interpret sensory data and execute precise physical actions in dynamic environments.

Parada said the focus on embodied intelligence allows AI to learn through physical experience rather than relying solely on static datasets [1]. This approach seeks to bridge the gap between large language models and physical actuators, the components that allow a robot to move.

The Tokyo Humanoids Summit served as the backdrop for these revelations, emphasizing the industry's push toward humanoid forms that can operate in human-centric spaces [1].

“Embodied intelligence is the next frontier of artificial intelligence.”

The transition toward embodied intelligence suggests a strategic pivot for AI labs, moving beyond chatbots and generative art toward physical automation. By developing models like Gemini Robotics-ER, Google DeepMind is attempting to solve the 'sim-to-real' gap, where AI performs well in simulation but fails in the physical world. Success in dexterous tasks like origami would signal a breakthrough in robotic manipulation, potentially accelerating the deployment of humanoid robots in logistics and domestic care.