Safeworld, a new startup led by Carnegie Mellon researchers, raised over $12 million to develop safety simulations for generative AI robots. The company aims to prevent accidents by testing robots in virtual environments with human models.
The company’s approach involves creating digital replicas of real-world scenarios, such as blind corners in factories, and simulating how robots would react to unpredictable human behavior. This method is similar to the testing processes used by companies like Tesla and Wayve, but adapted for the more complex and unstructured environments robots operate in.
"The time to build an industry safety standard is now while robots are being designed and deployed," said Jonathan Lai, a partner at a16z Speedrun. "By the time you have robots in households colliding with kids and causing safety incidents, that’s way too late."
Safeworld’s simulations will include a variety of human behaviors, such as tripping and falling, to ensure robots can respond appropriately in real-life situations. "Otherwise, you would have to go and trip and fall for the robot, which is like a hard thing to be doing all the time," said Kyle Wong, one of the company’s founders.
The company is still figuring out the best model for its product—whether to offer it as a platform for external users or as a services-based approach.
"We’ll probably be the first profitable company in this field," said Dr. Ding Zhao, the director of the Safe AI lab at Carnegie Mellon. "Because if anyone wants to deploy, they need to pay us to handle the situation."
Safeworld is partnering with companies like Gritt Robotics to develop safety simulations for their industrial robots.
"The difficulty with most of our systems is it’s very hard to formally prove it by doing some math, writing some equations, and saying yeah, the system is verified to be safe," said Vishal Dugar, the CTO of Gritt Robotics. "It necessarily has to be done empirically."
Source: techcrunch
