What tasks would you have a home robot do for you if you could?
There lies the crux of the issue.
For at least a decade, robotics and AI firms have flexed their ability to create machines that can complete various practical household tasks. Last month, Boston Dynamics showed its Spot robot dog picking up clothing. Miso Robotic’s Flippy has been flipping burgers for years. Other start-ups have demoed laundry-folding machines. The list goes on. More recently, news broke that Amazon may be closer to introducing a home robot capable of carrying things and following its owner’s commands.
But how likely is it that you’ll ever be able to own a true robotic butler?
Robots are indeed getting more complex. As AI continues to advance, it allows machines to figure out more complex problems and reliably chat with humans. Still, robotics and AI firms say you’ll have to wait quite some time before you’re able to own anything remotely similar to Rosey the Robot from “The Jetsons.”
In fact, companies are having a hard time commercializing anything more complex than a Roomba — which has been vacuuming houses for 20 years.
“We’ve been talking about home robots coming for a long time, and all we have so far is the vacuum cleaner,” said Jeff Burnstein, president of the Association for Advancing Automation. “You see demos, but taking it from a demo to something that actually works, then something you can afford, that’s the issue right now.”
Chore robots could take many forms. It could be humanoid, like Rosie. It might look more like a pet. No one really knows what might catch on. But we talked to some leaders in the space to find out what hurdles need to be solved before complex household robots could actually happen.
Right now, robots are doing well in factories where there’s plenty of space, no small kids around and employees wearing protective gear. They’re really good at completing a single repetitive task, like screwing on a wheel.
But imagine introducing machinery with legs and lifting capabilities into your home where things can and do go wrong. What if it falls on someone, or a software update causes it to go haywire? It’s funny on “The Jetsons,” but it wouldn’t be so comical if your grandmother were on the receiving end.
That's a significant challenge.
The biggest problem is safety, according to Marc Raibert, former CEO, and now chairman of Boston Dynamics, a robotics pioneer responsible for agile, animallike robots.
“The more complicated the robot, the more safety concerns. If you have a robot in close proximity to a person, and anything that goes wrong, that’s a risk to that person,” Raibert said. Things have gone wrong, at least on the job. In 2015, a 22-year-old man was killed while helping to set up a stationary robot at a Volkswagen plant in Germany. The robot pushed him against a metal plate and crushed him. In another case that year, a robot’s arm malfunctioned, hit and crushed a woman’s head in a Michigan auto plant.
It’s not that the safety issues at home can’t be solved. It’s that they haven’t been solved yet, robotics companies say. And making elaborate machines more household-friendly will almost certainly up the price.
Today’s mobile robots for factories can cost twice as much as an average new car. Take the robotic dog Spot, for example. It runs for about $75,000 without the arm attachment that makes it useful for transporting things. Without the arm, it’s basically a mobile surveillance machine. A humanoid robotic butler capable of autonomously completing a variety of tasks today could easily cost 10 times as much.
And robotics experts say people on the market for such a thing wouldn’t be willing to pay more than a few thousand dollars.
Anything higher than that, and it might be cheaper to pay a human to do the job. And humans just might do it better.
In 2020, Walmart pulled its inventory robots from the floor after reportedly finding that humans can scan products more simply and more efficiently than bulky six-foot-tall machines.
Robots also have a dexterity problem. Most can maybe pinch, grasp or use suction to hold an object. Meanwhile, humans with 10 fingers can manipulate things that come in various shapes and textures. Robotic limbs with humanlike flexibility do exist, but they cost tens of thousands of dollars to produce.
“The age of grasping is still a long way away. It is a very hard problem to solve at a mass consumption price point,” said Tim Enwall, CEO of Misty Robotics, a hardware company that builds personal robots for homes and offices.
Another challenge is that robots don’t know much about the world they’re operating in. You can teach them what an object looks like, but robotic butlers would also need to understand where in your home it’s located.
Take, for example, a simple task like taking a drink to the fridge. A machine would need to understand your command and be capable of autonomously navigating your home’s unique layout without spilling anything. It also needs to understand which room in your house is the kitchen.
That type of data requires collaborative input from consumers, and it’s something iRobot is working on by letting Roomba users mark where things are in their homes.
“The reason why iRobot doesn’t sell a robot with an arm is because we don’t know where anything is,” said Colin Angle, chairman and founder of iRobot. “We’re trying to get a better idea of what you actually want to do. That’s useful for your cleaning robot, but it’s imperative for whatever next robot you’re interested in buying.”
So when will something more advanced than a smart vacuum cleaner actually make an impact? Most say it’ll be decades from now.
Amazon is reportedly investing big in its “Vesta” home robot project, which may or may not see the light of day. (Amazon founder Jeff Bezos owns The Washington Post.) But robots will probably continue to crop up in new places as firms lay the groundwork for something more advanced to come along.
You’ll see more delivering food, carrying packages, cleaning surfaces and toting groceries. In other words, you’ll see more trying to perfect one single task.
“I do think we’ll find uses for home robots in the next eight to 10 years,” Raibert said. “Someone will pioneer a lower-cost mechanism that does a useful set of things. I do believe it; it’s just not tomorrow.”