The Moley cooking robot

Robots are playing an increasingly large role in restaurants, in fact they are expected to become mainstream there by 2025. Right now, the robot’s roll is mostly for mundane repetitive tasks, with the more artful bits of cooking still done by humans. There are a lot of reasons robots are relegated to menial tasks, and one of them is that it’s hard to program a robot to do something.

But a new startup spun out of UC Berkeley has developed technology that could bring a more human touch to robot cooking. Embodied Intelligence, which just raised $7 million in seed funding, is taking a new approach by having humans train robots through VR.

MIT Technology Review has a nice explainer on Embodied Intelligence, but for our purposes here, think of it this way: a human wearing virtual reality gear demonstrates how to perform an certain operation. The robot, through deep neural nets, learns how to mimic that VR operation.

As MIT Technology Review reports, more immediate capabilities “… will include picking complex shapes out of bins, assembling electronic components, and manipulating deformable objects such as wires, cables, or fabrics—skills that could translate well to advanced manufacturing settings.”

But it’s easy to imagine this technology expanding beyond the factory. Since the core of this process allows a robot to mimic a human, what if that human is a world-renown chef (or even just a competent one?). It’s fascinating to think of a master’s movements when it comes to slicing, or presenting a plate being precisely repeated.

True, a bucket of chanterelle’s is different than a bin full of bolts, but image recognition, AI and robot precision is only going to get better. It’s not hard to imagine a near future where robots move beyond flipping burgers and into more artful food preparation.

Of course, this also brings up the idea of soul in cooking. Can a machine just mimicking the actions of a chef could fully replicate the touch and experience a human puts into their creations? I don’t know, but if this technology works as advertised, we may not have to wait too long to find out.