Illustration by Andrew Nye for The Washington Post
The next generation of home robots will be more capable — and perhaps more social
Tech titans and upstarts alike are working to deepen our relationships with their robots
Wander around any big-box electronics store long enough and you may eventually stumble over a handful of domestic robots designed to tidy up after you. They’re probably somewhere near the refrigerators and the washers and dryers, waiting for you to introduce them to a mess.
That’s not exactly the future that decades of speculative fiction foretold. We were promised homes full of intelligent (and sometimes sassy) robotic assistants, not just the cavalcade of glorified, rolling appliances we wound up with. But over the next few years, a new wave of domestic robots could start to do more than just clean up after us.
Amazon introduced Astro earlier this fall, a $1,000-plus robot meant to ferry around small items and keep its eyes — well, cameras — peeled for intruders while roaming our homes. (Amazon founder Jeff Bezos owns The Washington Post.) Samsung, which at one point or another has built almost every kind of home gadget imaginable, has spent the last few years openly wondering about what a new generation of domestic robots should be able to do. And researchers and start-ups have continued to work on robots meant to help people in still more personal and social ways.
So, does this mean we’re on the verge of a home-robot renaissance? Maybe, but it’s not going to materialize overnight. Turning a more sophisticated breed of robots into actual products isn’t easy, which means the resulting machines won’t come cheap — at first, anyway. And long-standing concerns over privacy could dampen people’s enthusiasm for machines designed to recognize and react to members of their family.
That’s enough to convince Gartner analyst Annette Jump that these kinds of personal robots are “still six to eight years away from early mainstream adoption.” Until then, here’s what you need to know about how these machines are changing.
Cleaning your floors. Mowing your lawns. Scrubbing your grills. Most domestic robots you can go buy right now are single-purpose machines, and with good reason: Relatively speaking, robot vacuums like the Roombas are much easier to build than, say, Rosey from “The Jetsons.”
“That has really been the technique that has worked for a lot of different robots,” said Henny Admoni, the A. Nico Habermann Assistant Professor at Carnegie Mellon University’s Robotics Institute. “The Roomba does one thing and it does it well, and it’s the longest-lasting home robot we’ve seen.”
Slowly but surely, though, some of the biggest names in tech have designed robots that are meant to play multiple roles in our homes.
Consider Amazon’s Astro: The telescoping camera in its neck makes it a home security sentinel, while the camera embedded into the tablet that acts as its face means it can pull double duty as a video-calling station. Astro can use those cameras to recognize individual people’s faces, too, which will allow it to deliver small items to your kids or your partner sitting in another room. (That is, as long as there aren’t any stairs involved.)
And since this is Amazon we’re talking about, Astro is also an Alexa speaker on wheels, delivering reminders and playing music wherever it goes.
During its launch event, Amazon has said it tested Astro for “hundreds of thousands of hours” with employees and plans to release it to would-be beta testers as part of its Day 1 program later this year. (Long story short, that means you can pay a discounted price to get early access to Astro and provide feedback to the team behind it.)
While its plans to launch an Astro competitor are a little more secretive, Samsung has also been crafting robots of its own in forms and shapes that sometimes invite disbelief. At this year’s CES trade show, the company highlighted a concept robot called Bot Handy, a roving tower with “eyes” built into its cylindrical body, right next to a single, segmented arm.
Imagine sitting down for dinner after a long day at work — with a little help from artificial intelligence. Bot Handy uses its arm to gingerly pour you a glass of wine. And when you’re done with your meal, the robot could help load your dishwasher while making sure to use only the amount of strength needed to handle your fine china.
The weirdness doesn’t end there, either. At that same trade show a year earlier, Samsung showed off another off-the-wall concept: a grapefruit-sized orb called Ballie designed to roll itself around people’s homes, controlling smart home appliances, recognizing voice commands and curling up next to pets.
Samsung declined to comment on the progress of its home robots, but we do know two things. First, neither Bot Handy nor Ballie have ever been sold to real people. And second, Samsung is trying to shorten the gap between the moment it develops a home robot concept and the moment it actually goes on sale. Brian Harms, research engineer at Samsung Research America, has said his team now aims to make those projects marketable in one to three years, instead of the three to five years it used to target.
When it comes to robotics, Samsung is best known for its smart vacuum cleaners, some of which sell for more than $1,000. Going off the robotics projects Samsung keeps showing off at trade shows, though, the company clearly expects sophisticated, multitasking machines to eventually find a place in our homes.
And for better or worse, big companies like Samsung and Amazon seem to have the best chance of making those kinds of robots a reality — even though it’s too early to tell how well any of them actually work.
“They can use some of their cash to make the robot at a consumer price point,” Admoni said. “And that will let them at least get into the market and maybe survive long enough to adapt and to develop better systems.”
That doesn’t mean big companies haven’t struggled to make their personal robots worth using. Japanese conglomerate SoftBank reportedly stopped producing its charming Pepper robots earlier this year. But that hasn’t stopped some smaller outfits from trying to make their mark anyway.
The rise of social robots
In some ways, Buddy — a robot developed by a French company called Blue Frog Robotics — has a lot in common with Amazon’s Astro. It can remind you of events saved on your calendar. It can place video calls. It can even roam around your house in search of intruders. But for Blue Frog founder Rodolphe Hasselvander, Buddy is different in one crucial way: It’s meant to be an “emotional companion” as much as it is a gadget.
“The key point for acceptability and mass adoption is based on the creation of an emotional link with people,” he said.
Hasselvander has been working on Buddy since 2015, and — despite a few ups and downs — this cutesy robot is nearly ready to get to work. Blue Frog says it recently inked a deal with France’s Ministry of National Education that would see 1,750 of its machines used to help homebound or hospitalized children remotely interact with their peers in classrooms, and the company is gearing up to release Buddy in the United States in the next few months.
Unlike machines that perform a single mechanical task like cleaning, Buddy is an example of what researchers call a “social” robot.
“Social robots are designed to engage with people more as a collaborative partner,” said Cynthia Breazeal, director of the Personal Robots Group at the MIT Media Lab. “As opposed to a tool that you use, you interact more in an interpersonal way to achieve tasks or goals or experiences.”
Amazon’s Astro has some characteristics that Breazeal believes could make it a social robot of sorts, but the company’s product page plays up the machine’s home monitoring and security feature before almost anything else.
In addition to the practical tasks Buddy can perform, Blue Frog has also tested it as a companion to the elderly — who responded well to the robot’s reaction to petting and its expressive tablet-face — and as an educational tool for children with autism. But in those early tests, it also became clear that its users wanted Buddy to be able to do more, a shortcoming the Blue Frog team is trying to address by opening the robot up to developers who want to create their own apps and experiences for it.
“People want to have longer, more meaningful, more interesting conversations with these technologies. They get frustrated when it’s too transactional,” Breazeal said. “I think there’s a hunger and a desire for people to be able to interact with these technologies in this way.”
She would know. In addition to being a pioneer in social robot research, Breazeal also created Jibo, a gyrating bobblehead of a home companion that holds the Guinness World Record for being the most crowdfunded social robot. It developed a cult following for the sheer amount of personality it could be coaxed into expressing, but people’s Jibos collectively died an untimely death in 2019 when the servers that powered it were shut down. Jibo’s $899 price tag was hard for many to swallow and Amazon had announced its virtual assistant Alexa while Jibo was in development — that made it all but impossible for the robot to sustain a business.
(Don’t worry though: Jibo survives as a research tool and as a part of SoftBank’s robot division.)
Still, Jibo’s social legacy seems to live on in domestic machines like Buddy, even though some are wary of the baggage people might be carrying when interacting with it. The way robots behave in popular culture staples like “Star Trek” have given people a set of high expectations about what social robots should be able to do, and “the risk of putting a face or other anthropomorphic features on robots is queuing those expectations and potentially setting the bar too high,” Breazeal said.
And therein lies the rub for social robots: Because they’re meant to build fruitful, collaborative relationships with people, it can be difficult to nail down exactly how such a machine should look and act. Privacy remains a crucial concern for all home robots, but that’s especially true for social robots. As we allow those kinds of machines to act as our partners and companions, the potential fallout from hacks and improper data handling and analysis grows ever higher.
For those reasons, Admoni expects it will take much longer for social robots to catch on compared to their more immediately practical counterparts.
“I don’t think it’ll happen in the next couple years,” she said. “We haven’t yet had the breakthrough that will make the social robot we all imagined we could have in our homes. But I would be happy to be proved wrong.”