news 
Google DeepMind unveils new AI models for controlling robots
In a major breakthrough, Google’s AI research lab, DeepMind, has announced the development of new AI models called Gemini Robotics designed to enable real-world machines to interact with objects, navigate environments, and perform various tasks. These advanced AI models have been trained to generalize behavior across different robotics hardware and connect items that robots can “see” with actions they might take.
As part of this announcement, DeepMind has published a series of demo videos showcasing robots equipped with Gemini Robotics performing complex tasks such as folding paper, putting a pair of glasses into a case, and other actions in response to voice commands. This technology has the potential to revolutionize the field of robotics by enabling machines to operate independently in various environments.
The new AI models have been designed to allow robots to perform well in environments not included in their training data, which is a significant advancement over existing AI solutions. The lab claims that this technology can be used for a wide range of applications, including search and rescue operations, healthcare assistance, and more.
To facilitate the development of these advanced AI models, DeepMind has also released a slimmed-down version called Gemini Robotics-ER, which researchers can use to train their own models for robotics control. Additionally, the lab has created a benchmark called Asimov, which will help gauge the risks associated with AI-powered robots in real-world scenarios.
The implications of this technology are far-reaching and have the potential to transform industries such as manufacturing, logistics, and healthcare. The ability to create autonomous systems that can interact with their environment and make decisions independently could lead to significant improvements in efficiency and productivity.
However, it remains to be seen how quickly these advancements will translate into real-world applications. It’s possible that regulatory hurdles or societal concerns may slow the adoption of this technology. Nevertheless, the potential benefits are undeniable, and it will be interesting to see how Google DeepMind chooses to use their new AI models in the future.
Source: https://techcrunch.com/2025/03/12/google-deepmind-unveils-new-ai-models-for-controlling-robots/
