As robotization rapidly progresses and artificial intelligence (AI) continues to improve, tech giants are unveiling innovations that further change the current landscape. Recently, Google DeepMind introduced advanced AI models designed to enhance robotic control and their ability to adapt to changing environments.
The Latest News
On March 12, Google DeepMind, in collaboration with the Google AI research lab, announced its latest AI models called Gemini Robotics – an advanced vision-language-action model. These are designed for real-world machines to interact with objects, navigate environments, and more.
The company released a series of demo videos showing robots, using Gemini Robotics, performing tasks like folding paper, placing glasses in a case, and responding to voice commands.
According to DeepMind, this development brings creators one step closer to general-purpose robots.
This is because AI models need to exhibit three core traits:
- They must be general-purpose – able to adapt to various situations.
- They must be interactive – able to understand and quickly respond to commands or environmental changes.
- They must be dexterous – able to perform tasks typically done by humans using hands and fingers.
Previously, the company had advanced each of these areas separately, but now the goal is to merge them into one unified system that significantly boosts performance.
What Makes These Robots Unique?
According to DeepMind, during tests, robots operated in environments that were not included in the training data, demonstrating a new phase in their ability to adapt to entirely dynamic settings.
Alongside this announcement, the company introduced a simplified model, Gemini Robotics-ER, with advanced spatial understanding that allows developers to create custom robotic control programs.
It enhances the existing capabilities of Gemini 2.0 by combining spatial reasoning with coding skills. For example, upon seeing a coffee cup, the model can understand how to grasp it with two fingers, pick it up by the handle, and approach it safely along a correct trajectory. Researchers can use it to train their own robotic control models.
The Potential Impact of These AI Models on Automation
Both of these DeepMind’s models could significantly impact automation. The latest advancements allow robots to perform complex tasks and navigate dynamic environments, even when unfamiliar to the technology.
In the future, robots like Gemini Robotics could become more independent and widely integrated into industries. This will reduce the need for human intervention, automate more tasks, and enable them to perform work accurately and efficiently.
Final Thoughts
This is undoubtedly a major breakthrough in AI and robotics, bringing a greater potential for robots to become autonomous.
However, with these DeepMind’s AI models opportunities and the growing accessibility of such solutions, we will see changes in the job market and how industries are transformed by the possibilities of automation.
If you are interested in this topic, we suggest you check our articles:
- Google Gemini: How has it been received by users so far?
- Google’s Gemini 2.0 Flash Thinking Update
- Google’s Gemini Robotics and Gemini Robotics-ER: Intelligent Machines That Adapt and Learn
Sources: Google, TechCrunch, The Verge
