Safeworld is a new startup focused on using simulation technology to evaluate the safety of generative AI-powered robots in unstructured environments.
Key facts
- •Safeworld raised more than $12 million in a seed round led by Shine Capital and a16z Speedrun.
- •The company uses simulation software to test how robots interact with human models in unstructured environments.
- •Dr. Ding Zhao, who directs the Safe AI lab at Carnegie Mellon University, co-founded the company to address the unpredictability of generative AI.
- •Gritt Robotics is partnering with Safeworld to test its robotic systems used in industrial-scale solar farm construction.
- •The startup is currently evaluating whether to offer its safety verification as a platform for external users or as a service-based model.
Safeworld, a startup founded by Carnegie Mellon University's Dr. Ding Zhao, veteran executive Kyle Wong, and engineer Simo Rachidi, has emerged from stealth with over $12 million in seed funding. The company aims to address the safety challenges posed by generative AI in robotics by providing third-party validation for robotic control systems. The funding round was led by Shine Capital and a16z Speedrun, with participation from several other investors including the Carnegie Mellon University Endowment.
Simulating Human-Robot Interaction
Safeworld’s approach involves creating digital versions of physical environments, such as factory floors, using platforms like Genesis or MuJoCo. The company then inserts a simulation of the robot—driven by its actual software—to run thousands of scenarios where the machine encounters realistic human models. This process is designed to test how robots respond to unpredictable human behaviors, such as tripping, falling, or navigating around blind corners, which are difficult to account for with traditional algorithms.
Industry Need for Third-Party Validation
The founders argue that robot manufacturers require independent validation to ensure their systems are safe for deployment at scale. By acting as a third party, Safeworld intends to help competitors share safety information and establish industry standards. Gritt Robotics, which develops AI for robots used in solar farm construction, is currently partnering with Safeworld to empirically verify that its robotic arms can safely operate alongside human workers in various configurations and conditions.
Advertisement
This article was independently rewritten by ManyPress editorial AI from reporting originally published by TechCrunch.
