Nvidia announced Monday that it plans to release a software and hardware platform that can be used to build human-like robots that continue to learn through generative artificial intelligence (AI).
The platform consists of a computer system that powers the robot and AI, and a software suite that includes various tools such as genAI for building human-like robots.
Nvidia made this announcement at its annual developer conference, touting the capabilities that genAI provides to robots.
Jim Fan, research manager and head of embodied AI at Nvidia, posted on X that GR00T will allow robots to understand instructions and perform a variety of tasks through language, video, and demonstrations.
US-funded report issues urgent AI warning about 'out of control' systems attacking humans
“We are working with many leading humanoid companies around the world to ensure that GR00T can move into various embodiments and help the ecosystem thrive,” Fan said.
He also said that Project GR00T is the “foundation” of GEAR Lab's “Foundation Agent” roadmap. Huang said GEAR's team is building robots that learn to act skillfully in many worlds, both virtual and real. In his post, he also provided a video of team members working with the robot.
“These smarter, faster, better robots will be introduced into the world's heavy industry,” Reverend Reverendian, vice president of Omniverse and Simulation Technologies, told reporters. “We are collaborating with the global robotics and simulation ecosystem to accelerate development and deployment.”
Nvidia's “Jetson Thor” is the computer behind the genAI software, and the software package is called the “Isaac” platform.
Jetson Thor provides enough horsepower to enable robots to compute and perform complex tasks, while also allowing them to interact with other machines and humans, the company said.
What is artificial intelligence (AI)?
Isaac and new GR00T capabilities will enable “any robot in any environment,” Nvidia added.
Over time, the tool trains the software to improve decision-making through reinforcement learning.
In addition to the software and hardware platform, Nvidia announced that it will release pre-trained robot models and additional software to improve the sensing capabilities of the camera and the functionality of the robot arm.
Earlier this month, Nvidia CEO Jensen Huang announced that artificial general intelligence (AGI) could arrive in as little as five years.
NVIDIA CEO Jensen Huang says AI can pass most human tests within five years
The company's artificial intelligence chips are used to build systems such as OpenAI's ChatGPT, and Huang said the emergence of AGI will largely depend on how AGI's goals are defined.
“If I gave an AI… every test you can imagine, I would create a list of those tests and put it in front of the computer science industry, and in five years it would be successful. We'll go. One by one,” Huang said.
AI is currently able to pass exams such as the bar exam, but struggles with specialized medical exams such as gastroenterology.
NVIDIA CEO becomes $9.6 billion richer as AI chip maker's stock soars
However, the speed at which AI is developing and learning has many people worried that robots will take over the world.
The U.S. government commissioned a report released earlier this month calling for a “clear and urgent need to take action,” as the rapid development of artificial intelligence could lead to the extinction of humanity through weaponization and loss of control. ”.
“Rapidly expanding AI capabilities pose increasing risks to national security posed by weaponization and loss of control, especially given the fact that the continued proliferation of these capabilities amplifies both risks. There is clearly an urgent need to intervene,'' the report, published by Gladstone AI Inc., says.
CLICK HERE TO GET FOX BUSINESS ON THE GO
The report also recommended that the government strengthen controls over the production and export of AI chips.
Fox News Digital's Stephen Sorace and Eric Revell and Reuters contributed to this report.