Robots are a staple of the CES show floor. They’re fun to film, they grab attention, and, most importantly, they symbolize the futuristic fantasies that sustain so much of the tech industry. Without the dream of a robot butler cleaning your house, where would that next round of funding for the struggling robot startup come from?
But as with past shows, this year’s robots are a mixed bunch. More often than not, companies are touting functionality they can’t yet deliver, though there are some real trends hidden among the puffery. Let’s take a look at what the show has had to offer so far.
Robots that aren’t really robots
First up are the devices that are only borderline robots. These are things like the FoldiMate, a laundry-folding machine that we’ve seen before, but it actually works this year. There’s also the Bread Bot, a compact production line that turns flour, water, and yeast into freshly baked loaves of bread.
Is a dishwasher a robot? Probably not
You might want to call these robots (and that’s fine), but conceptually, they’re more like machines. They have a lot of moving parts and some sensing apparatus, but they don’t interact with the world around them. Famed roboticist Rodney Brooks defines a robot as something that can affect its surroundings, not just its interior. (It’s a good thing that the FoldiMate and Bread Bot aren’t able to practice their skills on the world at large.)
You might argue that, as with AI, we tend to call anything futuristic by its fancier title until the point that it becomes mundane. (If everyone had a laundry-folding machine in their house, it would just be another appliance, not a robot.) But it’s evident that these devices really are simpler than their truly robotic relations. That’s good news. It means they’re more likely to succeed, as long as they find their niche.
Robots that aren’t ready
Some robots look like they’re from the future, but they reveal that they’re stuck in the present when they’re asked to do anything. Usually, these are humanoid creations touted as robot butlers. Last year, for example, we had the Aeolus robot, which could sort of fetch you a beer from the fridge. But it was too slow, clunky, and prone to failure to be of much practical use.
So far, 2019 has been pretty quiet for this sort of device, though the bipedal Walker robot from Ubtech is probably the closest example. Ubtech is a big, well-funded Chinese firm that’s best known for its robot toys, but it also builds prototypes to showcase its advanced tech.
The Walker is such a robot. It stands nearly five feet tall, weighs 170 pounds, and can walk around and move its arms with the help of 36 actuators (robot muscles). It looks amazing, but it’s definitely more limited than it seems. Bipedal robots are extremely hard to build, and Ubtech has only demoed the Walker in predictable, controlled environments. The Walker makes for an impressive demo, but it’s a demo nonetheless.
Robots that are tablets on wheels
Sensing and navigation technology has gotten cheap and effective in recent years, as have voice interfaces. This means there are a lot of products touted as robots that are more like tablets on wheels. See, for example, Samsung’s Bot Care prototype, which the company presented as a way to look after the elderly, or the “personal butler” robot Temi, which stands 10 inches tall, is Alexa compatible, and has an Android tablet to show you information from the web.
Tablets on wheels are neat, but smart speakers are cheap
Devices like these have the potential to be useful, especially when it comes to communication with family or controlling smart home devices. But do they really have to toddle around on wheels? In lots of cases, this seems like a pointless complication, especially as smart speakers provide the same sort of on-demand computing access for a lot less work.
Robots that tug at the heartstrings
If there’s one thing robot research repeatedly shows, it’s that humans are extremely willing to bond with robots and treat them like fellow people. So, of course, the companies that make robots love to take advantage of this. At CES, the most prominent example of this trend was the Lovot, which was announced last December. This is a robot that’s designed to be loved and looked after — kind of like an advanced Furby with wheels and flippers.
Making a robot lovable isn’t necessarily a recipe for success. 2018 saw two high-profile social robots flop: Jibo and Kuri. But arguably, these devices couldn’t differentiate themselves from smart speakers in terms of functionality, while the Lovot is a more advanced bot, with sensors that detect touch, thermal cameras, and an impressive range of animated emotions. It looks sort of like a Pixar character, and it could be just as popular.
Robots that are really real
If I had to give an award to the robotiest robots of CES — those that fit the technical requirements but are also most likely to change your life in the near future — it would be the delivery bots. Over the last couple of years, this category has exploded. There are new startups, like Starship, Kiwibot, and Marble, while older companies like Domino’s and PepsiCo have also started experimenting with the technology. These robots didn’t have a huge presence at CES, but they were certainly rolling around, with new devices from relative unknowns like PuduTech as well-established players like Segway.
Delivery robots are low stakes and easy to make — they might just work out
It’s impossible to say for certain whether delivery robots will become a permanent fixture in our towns and cities, but they’re certainly a better bet than robot butlers. They take advantage of cheap sensors, stable navigation provided by recent AI advances, and they’re not as high stakes as self-driving cars. (If they mess up, they don’t kill pedestrians.) They work on pavements but also inside buildings, and they meet the needs of a population that increasingly shops online and expects delivery on-demand.
If there’s one thing to take away from the robots at CES — and the state of the consumer robotic industry more generally — it’s that it pays to specialize. The technology to build the do-it-all robot of our dreams just isn’t available and likely won’t be for a while. As we’ve seen with robot vacuums, if you want your device to succeed, it has to find a niche that works within the limits of current hardware. For a robot to succeed, it has to be useful or, failing that, lovable. Robots that are neither won’t get far.