A team of scientists led by Northwestern University has recently devised an artificial intelligence (AI) program that has the groundbreaking capability to design walking robots from scratch.
To demonstrate the AI’s prowess, the researchers provided it with a seemingly elementary instruction: “Design a robot that can walk across land.” While nature required billions of years to bring forth a walking species, this new technology astoundingly achieved the objective within seconds.
However, the AI’s uniqueness is not solely in its speed. Unlike many of its counterparts, this system operates on a standard personal computer, conceptualizing entirely new designs.
Most other AI models rely on vast datasets processed on high-powered supercomputers. Despite these resources, they’re largely limited by humans’ previous instructions, often reproducing past human designs instead of generating innovative ones.
“We discovered a very fast AI-driven design algorithm that bypasses the traffic jams of evolution, without falling back on the bias of human designers,” said co-author Sam Kriegman, an assistant professor of Robotics and AI at Northwestern.
“We told the AI that we wanted a robot that could walk across land. Then we simply pressed a button and presto! It generated a blueprint for a robot in the blink of an eye that looks nothing like any animal that has ever walked the earth. I call this process ‘instant evolution.’”
Previously, in an early phase of the research in 2020, Kriegman had already been remarked by the scientific community with his development of xenobots, a pioneering type of robots crafted entirely from biological cells.
Now, Kriegman and his team consider the recently developed AI an evolutionary leap, coming closer to harnessing the untapped potential of artificial life.
“When people look at this robot, they might see a useless gadget. I see the birth of a brand-new organism,” Kriegman said.
Highlighting the AI’s powerful adaptability, the experts began with the foundational task of designing a machine fit for terrestrial movement. Although the AI began with a simplistic algorithm, through iterative advancements, it continually improved upon its initial design.
Astonishingly, in just nine design cycles, the AI crafted a robot capable of walking at a pace nearly equivalent to half of an average human’s stride.
All these design stages, right from the inception of a rudimentary block to a fully functional walking robot, were executed in a breathtaking 26 seconds on an ordinary laptop.
Reflecting on this extraordinary achievement, Kriegman said: “Now anyone can watch evolution in action as AI generates better and better robot bodies in real time. This is because evolution has no foresight. It cannot see into the future to know if a specific mutation will be beneficial or catastrophic. We found a way to remove this blindfold, thereby compressing billions of years of evolution into an instant.”
Amazingly, without any human input, the AI conceived legs as the answer to movement on land. However, its design deviated from nature’s symmetrical templates.
“It’s interesting because we didn’t tell the AI that a robot should have legs,” Kriegman said. “It rediscovered that legs are a good way to move around on land. Legged locomotion is, in fact, the most efficient form of terrestrial movement.”
Exploring the tangible implementation of the AI’s simulation, the team used the robot’s design as a blueprint, transitioning it from the virtual realm into the physical world using 3D printing and silicone rubber.
While the robot’s ability to walk is an undeniable marvel, Kriegman envisions a future teeming with diverse applications for tools born out of this AI’s ingenuity.
Elaborating on potential use-cases, he suggested scenarios where such robots could navigate disaster sites, seeking out survivors, or delve into intricate networks like sewage systems for maintenance and repair. On a microscopic scale, AI-designed nanobots might one day journey through human bloodstreams, providing medical interventions.
“The only thing standing in our way of these new tools and therapies is that we have no idea how to design them. Lucky for us, AI has ideas of its own,” Kriegman concluded.
The study is published in the journal Proceedings of the National Academy of Sciences.
image Credit: Northwestern University
Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.