Next-Gen AI Robots? Google’s Gemini Models Aim for Real-World Intelligence

Tech giant Google on Wednesday said it will bring its DeepMind artificial intelligence technology models to power robots. The company, in a blog post, stated that its latest AI models --- Gemini Robotics and Gemini Robotics-ER --- will run on Gemini 2.0. According to Google, Gemini 2.0 is the “most capable AI” till the date.

Gemini Robotis is an advanced vision-language-action (VLA) model built on Gemini 2.0, with physical actions as a new output modality for controlling robots directly. On the other hand, Gemini Robotics-ER enables roboticists to run their own programs using Gemini’s embodied reasoning abilities.

“Both of these models enable a variety of robots to perform a wider range of real-world tasks than ever before,” the blog read.

For this, the company will collaborate with Apptronik, a Texas-based robotics developer, to build the next generation of humanoid robots with Gemini 2.0, it said. The developer has previously worked with Nvidia and NASA. Earlier, Google also joined its $350 million funding round.

Read More | Manus AI: China's New Artificial Intelligence Agent Has Its Own Mind

AI’s Key Qualities for Robotics

For robotics, AI models need three essential qualities like general, interactive, and dexterous. “They should be able to adapt to different situations, they should understand and respond quickly to instructions or changes in their environment, and the third one, they can do the kinds of things people generally can do with their hands and fingers like carefully manipulate objects,” said Google.

“This kind of control, or “steerability,” can better help people collaborate with robot assistants in a range of settings, from home to the workplace. We trained the model primarily on data from the bi-arm robotic platform, ALOHA 2, but we also demonstrated that it could control a bi-arm platform, based on the Franka arms used in many academic labs,” it added.

Read More | OpenAI Eyes 'PhD-level' AI Agents Priced At $20,000 Per Month: What Will It do?

Sundar Pichai on Robotics

In a post on X (formerly Twitter), Google CEO Sundar Pichai said the company sees robotics as a helpful testing ground for translating AI advances in the physical world.

“Today we’re taking our next step in this journey with our newest Gemini 2.0 robotics models. They show state of the art performance on two important benchmarks - generalisation and embodied reasoning - which enable robots to draw from Gemini’s multimodal understanding of the world to make changes on the fly + adapt to their surroundings,” he said.

“This milestone lays the foundation for the next generation of robotics that can be helpful across a range of applications,” Pichai added.