Illustration by Toby Leigh of a robot babysitter
© Toby Leigh

If ever you start to fret about being replaced by a robot in some Terminator-style apocalypse, look up a YouTube video entitled “PR2 Autonomously Pairing Socks”. The two-minute clip from a University of California, Berkeley research project in 2011 shows a robot identifying the two matching pairs from five socks. After carefully flattening the socks to inspect their shape and pattern, it checks they’re not inside out, then couples them.

It’s an impressive feat of robot engineering, of course. But for some it will be quite comforting that this task, dispatched in seconds by us humans, takes the $280,000 Willow Garage PR2 robot fully half an hour. On a cost-benefit basis alone, robot armies are unlikely to be marshalled against us in the near future.

Yet robots are about to invade our homes – and they’re coming not for our laundry but for our children. Rather than androids with arms, clunky thumbs and screens for eyes, they’re arriving as toy cars, balls and babysitters – and costing not $200,000 but $200. You might call them robots in disguise.

One company in the vanguard of the rise of the friendly robots is Anki, a start-up that launched onstage at this month’s Apple developer conference in San Francisco, after raising $50m. Boris Sofman, co-founder and chief executive, raced four small cars around a special mat, controlled over Bluetooth by an iPhone. So far, so 21st-century Scalextric. But, like a video game brought to life, none of the cars was being driven by a person. Each was given its own mission – to overtake another car, say – and used artificial intelligence to try to beat its opponent. The toys use a camera sensor to position themselves on the track, make thousands of calculations a second to decide what to do, then act on that reasoning.

Sofman and his fellow roboticists used their experience building much larger autonomous vehicles at Carnegie Mellon University to create Anki Drive. Drive will be the first Anki product to go on sale this autumn. A set-up that cost tens of thousands of dollars in the robotics lab is now being recreated with a few dollars’ worth of sensors, cameras and chips, with most of the processing done on a smartphone.

“The hard part is not just solving the robotics problems but solving them reliably at a price point that makes a consumer product,” Sofman says. “We are taking these projects that were funded by Darpa, Nasa, John Deere and Caterpillar and launching a game that somebody can buy for Christmas for their kids.”

Sofman predicts that component prices will continue to fall, leaving no area of our lives untouched by robots and artificial intelligence. Marc Andreessen, the noted Silicon Valley investor, is backing Anki, calling it “the best robotics start-up I have ever seen”.

Other promising robo-toys include Sphero, a $130 self-propelling ball controlled by a smartphone. Then there’s Romo, a little tractor with an iPhone for a face (and brain). Romo acts as a roving video camera and speaker for the person holding the connected iDevice, whether in the same room or across the world. An employee at Romotive, Romo’s developer, told me she’d used it to babysit her young niece from across the country while her mum had a nap.

What’s exciting about these robots is that they don’t just live in the future or a lab; they are available today, to anyone with a smartphone or tablet. As the emphasis shifts from standalone apps to fitness trackers and WiFi lightbulbs, the smartphone revolution is providing a platform for a whole new world of connected devices.

Robots, previously banished to drudgery in the manufacturing or defence industries, are heading for the living room. And while one child with an iPhone can be playing with, say, an Anki car, an adult with another robo-connected smartphone might be babysitting him from 1,000 miles away. Now that is kind of scary.

Tim Bradshaw is the FT’s San Francisco correspondent

Copyright The Financial Times Limited 2024. All rights reserved.
Reuse this content (opens in new window) CommentsJump to comments section

Follow the topics in this article

Comments