DeepMind to Harness AI Models for Enhancing Physical Robots
Google DeepMind Unveils New Robotics AI Models
Google DeepMind recently introduced two innovative AI models aimed at revolutionizing the robotics field. These models, named Gemini Robotics and Gemini Robotics-ER (Extended Reasoning), operate on Gemini 2.0, which Google touts as its "most capable" AI system to date. Unlike previous AI models, which primarily focused on generating text and images, the new Gemini models enable real-world actions by controlling robots.
Partnership with Apptronik
In a strategic move, Google has partnered with Apptronik, a Texas-based robotics developer, to enhance the development of humanoid robots using the Gemini 2.0 technology. Apptronik is known for its collaborations with prominent organizations, including NVIDIA and NASA. Recently, Google participated in a $350 million funding round for Apptronik, showcasing their commitment to advancing robotics together.
Demonstration of Capabilities
During the announcement, Google showcased videos of Apptronik robots performing various tasks using the new AI models. The demonstrations included actions like plugging devices into power strips, filling lunchboxes, arranging plastic vegetables, and zipping bags, all in response to spoken commands. Although the exact timeline for the public release of this technology remains unclear, these demonstrations highlight the practical applications of robots powered by Gemini.
Essential Qualities for Robotics AI
Google emphasizes three key qualities that AI models for robotics must possess:
- General Adaptability: The ability to perform in a variety of situations.
- Interactivity: Understanding and quickly responding to spoken instructions or environmental changes.
- Dexterity: The capability to manipulate objects with precision, similar to human hand movements.
These qualities are crucial for creating robots that can be genuinely useful and supportive in daily tasks.
Offering Gemini Robotics-ER
The Gemini Robotics-ER model offers a foundation for roboticists to develop their own AI models. It will be made available not only to Apptronik but also to a select group of "trusted testers," including notable companies like Agile Robots, Agility Robotics, Boston Dynamics, and Enchanted Tools. This initiative aims to foster innovation in robotics by allowing developers to experiment with and enhance the Gemini technology.
Competing in the Robotics AI Landscape
Google is not alone in the race to integrate AI into robotics. OpenAI recently invested in Physical Intelligence, a startup focused on creating AI that can function effectively in the physical world. This investment underscores the increasing interest and competition in the realm of robotics powered by AI.
Additionally, OpenAI has made strategic hires to bolster its robotics initiatives. Tesla, too, has ventured into the humanoid robotics space with its development of the Optimus robot, further intensifying the competition among tech giants in this sector.
CEO Insights on Robotics and AI
Google’s CEO, Sundar Pichai, expressed that the company views robotics as an ideal testing platform to translate AI advancements into real-world applications. He stated that the robots are designed to leverage Google’s multimodal AI, enabling them to adapt to various environments and make real-time adjustments.
As advancements in AI and robotics continue to evolve, the potential for creating intelligent, interactive robots becomes more promising, potentially transforming industries and everyday life.
