The stunningly obvious, invariably overlooked case against all these robot companions

Let's count the ways robots failed to deliver at CES 2018
Written by Greg Nichols, Contributing Writer

It wasn't a great showing for robots at this year's CES.

Things got off to a rocky start when CLOi, the ringleader for LG's new artificial intelligence-equipped appliances, failed repeatedly to follow voice commands from LG's vice president of marketing, David VanderWaal.

Sony's updated, Wi-Fi enabled Aibo, perhaps in cross-brand solidarity with CLOi, later ignored CEO Kaz Hirai during a Sony press event.

FoldiMate, a clothes folding robot that costs nearly $1000, raised some eyebrows, but not for the right reasons. Attendees were underwhelmed to learn that clothes have to be fed in one at a time, somewhat eliminating the time-savings the device is supposed to deliver. It also reportedly jams easily.

Aeolus, one of the conference favorites, is a latter day Rosie the Robot. It got high marks from a lot of tech reporters for having a dexterous gripper, enabling it, say, to fetch a beer or hold a vacuum. But the demos were constrained to a few pre-programmed use cases in a known environment, and the unit will likely cost in the high thousands. It's a hint at what a home robot can be, but it isn't ready for prime time.

There were plenty of other flat notes. A live event featuring brand new tech is bound to yield some bloopers. But the robots at this year's CES failed in a way that's less easy to overlook: They failed to make a case that we need robots in the home right now--or at least that we need any new robots in the home.

To understand why, you need only consider what automation is for. In a nutshell, it's supposed to save humans time. That's really it. That's what robots are all about.

Yet most of the companies presenting robots at CES seem to have set aside the notion of helpful automation in favor of a less quantifiable, more abstract sales pitch, one built around engagement, interactivity, and obeisance to technological complexity.

There's a good reason for that. Turns out automating chores is really tough.

There are all kinds of activities we do grudgingly in the course of a day. We put groceries away, cook dinner, make our beds, sort the laundry, bathe our pets. But for as far as robots have come in terms of dexterity, navigation, and sensing capabilities, there really aren't that many physical tasks inside the home that the current spate of personal robots comes anywhere close to being able to accomplish reliably.

The small number of chores that can be easily automated already have been. Two of the biggest time sucks in generations past were solved by the most overlooked robots in modern life: washing machines and dishwashers. It's hard to imagine how whiz-bang crazy those technologies seemed when they were first introduced.

More recently, iRobot and a long list of other companies have tackled the ho hum task of vacuuming contiguous rooms not separated by stairs. By the same token, in Western Europe, where cheap human labor is less readily available than in the U.S., lawn mowing robots have gained popularity.

There are a few other examples. Refilling ice trays is a pain, so modern freezers do it for you. Keeping track of the temperature is burdensome, so a thermostat in your house flips the switch on your heater or AC without bothering you.

But putting the dishes away? Cooking a meal from raw ingredients? We're more than a decade out from that kind of automation, at least in a marketable appliance. Today's industrial robots are certainly dexterous enough for those tasks, but they're not adaptable to the varied landscapes of people's homes. They're also wildly expensive in the context of consumer electronics. And there's the issue of form factor. No one wants a KUKA-orange appendage protruding from their kitchen floor.

A clothes folder may be the next big thing for home robotics. Folding the laundry is time consuming and monotonous, and the deviations between sets of clothing can be accounted for predictably. But again, the FoldiMate isn't quite there. Maybe in a few years.

In the meantime, the home robots that are being rushed to market today are little more than rolling Amazon Echoes. In fact, that's the market personal robotics developers seem keenest to capture. It's a big one, to be sure; something like one-in-six Americans now has an AI-powered assistant.

But the success of tabletop personal assistants doesn't suggest that mobile, embodied AI assistants are the next big thing. That one-in-six figure is primarily the result of the unique ability of Amazon and Google to offer hardware for next to nothing, coupled with a genuine need, which has arisen in the last few years, to take fingers out of the process of performing a web query, playing music, or changing a setting in your smart home while you're otherwise engaged. Amazon and Google have solved that problem beautifully.

For now, slapping wheels on Alexa and giving Google a face are costly extras. Those features may be cool, and there's a case to be made that they enhance user experience, but at root they're features that don't solve a core problem for most consumers.

That's not to say the day won't come when Rosie the Robot will clean your house and cook dinner--clearly that's where we're headed.

But if home robots are going to be taken seriously, robotics companies need to start thinking less about rushing to get new robots into the home and more about automating the mundane tasks we'd all rather not do. That's the real promise of a robotic future.

Editorial standards