DeepMind Announces Plans to Utilize AI Models for Enhancing Physical Robots
Google DeepMind Unveils New AI Models for Robotics
Google DeepMind has recently launched two pioneering AI models designed specifically for robotics, both powered by Gemini 2.0, which the company claims to be its most advanced AI yet. This initiative marks a significant step in integrating artificial intelligence into practical robotic applications.
Introducing Gemini Robotics
The new AI models are called Gemini Robotics and Gemini Robotics-ER (Extended Reasoning). Unlike traditional generative AI that focuses primarily on text and images, Gemini Robotics is designed to facilitate physical actions by controlling robots. This advancement aims to enhance robots’ ability to perform various tasks in real-world settings.
Partnership with Apptronik
To develop the next generation of humanoid robots, Google has formed a collaboration with Apptronik, a robotics firm located in Texas. Apptronik has previously collaborated with notable entities such as NVIDIA and NASA and recently announced Google’s involvement in a significant funding round, raising $350 million.
In various demonstration videos, robots equipped with the new AI models showcased their capabilities by performing tasks such as plugging items into power strips, filling lunch boxes, moving plastic vegetables, and zipping up bags—all in response to spoken commands. However, Google has not specified a timeline for releasing these technologies to the public.
Key Features of Robotics AI Models
Google emphasizes that for AI models to be effective in robotics, they need to possess three essential qualities:
- General Adaptability: The ability to adjust to different situations and environments.
- Interactivity: Understanding and responding promptly to instructions or changes around them.
- Dexterity: The capability to perform tasks similar to human hand movements, allowing for careful handling of objects.
Gemini Robotics-ER for Developers
The Gemini Robotics-ER model serves as a foundational tool for roboticists, enabling them to train their models effectively. This model is accessible to Apptronik and other "trusted testers," including renowned companies like Boston Dynamics and Agility Robotics.
Competitive Landscape in Robotics AI
Google is not the only company venturing into AI for robotics. OpenAI, for instance, has made substantial investments in Physical Intelligence, a startup dedicated to introducing general-purpose AI in the physical realm. OpenAI is also ramping up its robotics initiatives by hiring leadership from Meta’s augmented reality division to spearhead its consumer hardware and robotics projects.
Tesla, known for its innovative strides, is also exploring humanoid robotics through its Optimus robot.
Broader Implications of Robotics and AI
According to Google CEO Sundar Pichai, robotics can serve as an experimental ground for applying AI advancements in tangible environments. By leveraging multimodal AI models, these robots aim to adapt and make real-time changes based on their surroundings, ultimately enhancing their functionality and usability.
Investments in AI technologies like those from Google and OpenAI highlight the growing importance of robotics in everyday life. With ongoing advancements, the future possibilities for integrating AI into practical, human-like robots appear promising, potentially transforming various industries and boosting productivity across sectors.
