DeepMind Announces Plans to Utilize AI Models for Physical Robots
Google Unveils Advanced AI Models for Robotics
Introduction to Google’s New AI Models
On Wednesday, Google DeepMind introduced two groundbreaking AI models designed specifically for robotics. Operating on the Gemini 2.0 platform, these models—Gemini Robotics and Gemini Robotics-ER (Extended Reasoning)—mark a significant advancement in artificial intelligence. This new technology aims to transition AI capabilities from primarily generating text and images into the realm of physical actions, allowing robots to perform tasks more effectively.
Partnership with Apptronik
Google has announced a collaboration with Apptronik, a Texas-based developer specializing in robotics. This partnership is set to pioneer the creation of the next generation of humanoid robots powered by the Gemini 2.0 AI system. Apptronik has an impressive track record, having previously worked with notable organizations like Nvidia and NASA. Recently, Google participated in Apptronik’s $350 million funding round, further signaling its commitment to advancing robotic technology.
Demonstrations of Robotic Capabilities
In a series of demonstration videos, the capabilities of Apptronik robots equipped with the new AI models were showcased. These robots displayed a variety of tasks, including:
- Plugging devices into power strips
- Filling a lunchbox
- Moving plastic vegetables
- Zipping up bags
These activities were executed in response to spoken commands, showcasing the robots’ ability to understand and act quickly to instructions.
Core Qualities of the AI Models
Google outlined three essential characteristics that AI models for robotics must possess to be considered effective:
- Generality: The AI should be adaptable and able to handle diverse situations.
- Interactivity: The robot must be capable of understanding and responding to real-time changes and instructions.
- Dexterity: The ability to manipulate objects similarly to how humans do, using fingers and hands.
These features are crucial for ensuring that robots can perform tasks safely and efficiently in various environments.
Gemini Robotics-ER for Developers
The Gemini Robotics-ER model is tailored specifically for roboticists, providing a foundational framework for training and developing custom AI models. Apptronik and a select group of "trusted testers," which includes organizations like Agile Robots and Boston Dynamics, will have access to this model to enhance their robotic capabilities.
Broader AI Robotics Landscape
Google is not the only tech giant investing in AI for robotics. In a notable move, OpenAI announced an investment in a startup called Physical Intelligence, which focuses on integrating general-purpose AI into the physical world. OpenAI has also made strategic hires, including the former head of Meta’s initiative on augmented reality glasses, to spearhead its robotics efforts.
Furthermore, companies like Tesla are entering the humanoid robotics sector with products like the Optimus robot, illustrating the growing competition and interest in this field.
Google’s Vision for Robotics
Sundar Pichai, Google’s CEO, expressed that the company views robotics as a vital testing ground for applying AI advancements in tangible applications. He emphasized that these robotic systems would utilize Google’s multimodal AI capabilities to adapt dynamically to their environments, enhancing their efficacy in real-world scenarios.
Overall, the developments unveiled by Google in the realm of robotics herald a new era of intelligent machines capable of interacting with the physical world as efficiently as humans.
