AI-powered robot dog is learning to 'live' like a human

This is Luna, Intuacel's version of the robot dog

And while it may look like its predecessors, Luna has something they don't

A functioning digital nervous system that allows it to learn the way sentient beings do

Traditional AI learning models have two ways of acquiring the knowledge they need to perform tasks

One is to have humans program instructions that the machine can follow step by step

step. The second, which has been widely used by recent models, is called machine learning

and it requires feeding the machine a certain amount of data, which it can then use to draw

conclusions, make predictions, and decisions. It's what chat GPT and most currently popular AI

models are using. But Intuacel argues that Luna's model is something else. Its technology is called

physical agentic AI, and it allows the machine to make its own decisions and take independent

action towards specific goals without the need of pre-existing data. Instead, physical agentic AI

learns empirically, directly from the environment it occupies, in the same way that humans or other

sentient beings do. In the past, there have been models like OpenAI's spinoff Covariant and its

robotics foundation model, which use pre-existing data to build expertise that would help make

decisions while completing a task. But these kinds of models still need that pre-existing data

in order to build knowledge. And while some machines get their training from data Luna has its own certified dog trainer who usually works with real dogs but will now teach the robotic canine how to walk

In a series of tests run by N2SL, you can see Luna try to stand up on its feet just

like a newborn animal, learn to walk on difficult terrain, and attempt to balance on slippery ice

The purpose of all of this, according to Intuosel, is to create robots that could one day be sent to distant planets in space, like Mars

As Earth-based simulations can't prepare robots for all the challenging conditions they will be exposed to in space

the goal is to create machines that have the ability to learn from and adapt to the environment they're in

And while this little guy is still learning to get around, Intuosel says that the next step for them is to apply physical agentic AI to human-like robots

So what do you think? Is this an exciting new step in space exploration

Or are we going a little bit too far? I don't think you're giving us our due credit

Our scientists have done things which nobody's ever done before. Yeah, yeah, but your scientists were so preoccupied with whether or not they could that they didn't stop to think if they should

Let us know in the comments. And if you're interested in learning more about the future of AI and robotics, head over to Mashable.com