With clunky determination, robots played soccer, impressed children with shadowboxing, and shot arrows Monday at the historic site of the Olympic Games.
As the robots shuffled and sometimes paused for battery swaps, their developers and futurists discussed the key question: when will robots be capable of chores like cleaning closets and doing dishes?
Exploring Space Before Tackling House Chores
Despite rapid progress in AI applications like ChatGPT, humanoid robots—those with human-like looks and abilities—are still years behind.
“I truly believe humanoids will explore space first and then enter our homes… the home is the ultimate challenge,” said Minas Liarokapis, a Greek academic and startup founder who organized the International Humanoid Olympiad.
The four-day event brought together experts and developers at Ancient Olympia in southern Greece, the site where the flame is ignited every two years for the modern Summer and Winter Olympics.
“To enter homes will take over 10 years, definitely more,” Liarokapis said. “I mean performing tasks with skill, not just selling robots that are cute companions.”
Insufficient Training Resources
AI is advancing rapidly due to the abundance of online data. However, training resources for humanoid robots are limited because real-world actions are slower, costlier, and more difficult to capture than digital data such as text or images.
According to an article in the latest issue of Science Robotics, humanlike robots lag behind AI by about 100,000 years in data learning.
To close this gap, Ken Goldberg, a professor at the University of California, Berkeley, recommends that developers go beyond simulations and blend “traditional engineering” with hands-on training. He believes this approach would enable robots to “gather data while doing practical tasks, like driving taxis or sorting packages.”
The Quest for Valuable Data
Luis Sentis, professor of aerospace engineering and engineering mechanics at The University of Texas at Austin, emphasized that successful robotics depends on collaboration among researchers, data firms, and large manufacturers to achieve scale. He pointed out that these partnerships are already drawing billions in investment to advance humanoid robots.
“These collaborations are progressing rapidly, and I see these challenges being solved daily,” said Sentis, who is also a co-founder of the humanoid robotics company Apptronik.
At the Greek event, developers showcased their own innovations.
Aadeel Akhtar, CEO and founder of advanced prosthetics company Psyonic, gained worldwide recognition after pitching his company’s bionic hand with sensory feedback on the U.S. TV show “Shark Tank” last year.
He told The Associated Press on Monday that the data from this could speed up robot development.
“We designed our hand for both humans and robots,” he explained. “By using the prosthetic hand on humans and transferring that data to robots, we’re helping close the gap.”
Neurons
Hon Weng Chong, CEO of Cortical Labs, said the Australian biotech firm is working on a “biological computer” powered by living brain cells cultivated on a chip. These cells are capable of learning and reacting to information, potentially helping robots think and adapt more like humans.
At the Olympiad, organizers aimed to establish a foundation for yearly competitions that offer a “genuine assessment of the progress made in humanoid robotics,” said Patrick Jarvis, co-founder of robot company Acumino, alongside Liarokapis.
Organizers kept the events within the realistic capabilities of humanoid robots.
“We wanted to include discus and javelin, but those are challenging for humanoids,” Jarvis explained. “We also can’t judge a high jump because it would require specially built legs—which most humanoid robots don’t need.”
China is Eager to Showcase its Robots, While the U.S. Remains more Reserved
One company even experimented with having its robot attempt the shot put, according to Thomas Ryden, executive director of MassRobotics, who helped recruit as many humanoid companies to the event as possible.
Ultimately, while several American roboticists attended the event in Greece to give talks, few brought their robots.
In contrast, Chinese companies are more frequently presenting their robots at public events—like Beijing’s inaugural Humanoid Robot Games in August—while U.S. counterparts tend to rely on slick promotional videos that often hide technical shortcomings.
There are notable exceptions. In 2022, Elon Musk introduced Tesla’s Optimus: the prototype walked stiffly onto the stage, turned, and waved to an enthusiastic audience.
Boston Dynamics took things a step further. A decade after debuting its dog-like robot Spot, the company featured them dancing in sync to a Queen song on America’s Got Talent.
One of the five robots malfunctioned mid-performance, turning into a reality-show moment—but also demonstrating the robots’ impressive coordination and agility.
“Can I be honest with you? I actually think—and I don’t mean this harshly—it was strangely better that one of them failed,” said judge Simon Cowell. “Because it showed just how hard this is.”
Read the original article on: Tech Xplore
Read more: Why are We so Fascinated with Making Robots Dance?