Physical AI: Gecko Robotics wants humans controlling robots

Jake Loosararian, CEO of Gecko Robotics, has a pretty straightforward idea: if AI-driven robots are stepping into the real world, humans need to stay in the driver's seat. This isn't sci-fi; it's just common sense for autonomous systems we'll soon see everywhere. We're talking about machines that don't just think, but actually do things and interact with their environment.

And when these machines start moving, perhaps in a factory or, even worse, in military settings, someone needs to be able to hit the "stop" button. Loosararian told Bloomberg Tech that human control is crucial for physical AI safety. A simple concept, yet one often forgotten amidst all the AI hype.

Gecko isn't just sounding the alarm; they're offering a solution. They're working with Nvidia on an "open agent safety platform." Imagine a sort of digital fence that stops robots from getting too... creative.

This platform is designed to set clear boundaries for the AI agents controlling robots. Whether it's robotic arms in a factory or drones in the field, the idea is to establish precise guardrails. Gecko Robotics is developing an open safety platform with Nvidia for AI agents operating robots, intended for industrial and military sectors. Sounds a bit like a manners guide for machines, doesn't it?

Usually, we assume adding safety means hitting the brakes on progress. But Loosararian isn't buying it. He argues that safety and rapid tech deployment aren't necessarily enemies. In fact, they can and should be balanced right from the start.

For him, integrating safety from day one actually speeds up adoption. It's not about adding brakes later, but building with them as standard equipment. Jake Loosararian contends that safety and rapid implementation of physical AI don't have to be at odds; they can go hand-in-hand. An approach that, frankly, makes a lot of sense.