Household Robots: Why They Are Not Worth the Money Yet
Households robots like Jibo or Amazon’s Echo have arrived in the living rooms of tech enthusiasts. For now, they perform a lot of the same tasks that you could complete on your smartphone or a tablet connected to your home’s WiFi network. Experimenting with one reveals that these robots, fun though currently superfluous, have a lot of room for improvement, perhaps in the ways they recognize speech or retrieve the information you request of them.
But Dan Mitchell reports for MIT’s Technology Review that focusing on what household robots can do right now might be the wrong way to evaluate the value that they could potentially bring to a home. Instead, he suggests, imagine what they might be able to do in five years, ten, or fifty — highlighting the fact that the trend toward domestic robots is only in its early stages. So early that there’s still debate over a key term in the area, “social robot,” and what it really means.
Both Jibo, a robot that can greet you, remind you of tasks, and deliver messages to different members of the household, and Amazon Echo, which you can ask for information, music, news, or weather, are both referred to as social robots. But as Mitchell notes, there are big differences between the two. He explains that Jibo is animated and highly interactive — much closer to what people think of when they think “robot” than the monolithic Echo, which primarily responds to simple commands. Jibo supports video, while Echo does not. (And Jibo costs $749, versus Echo’s $199.)
Though Jibo can move, neither it nor Amazon’s Echo is mobile. And there’s not much reason for either of them to be; domestic robots can’t yet do household work like whipping up a snack or tidying the family room. Cynthia Matuszek, a robotics researcher at the University of Maryland, Baltimore tells Technology Review that robots will eventually be able to take on tasks like cooking gumbo, but like most roboticists, she doesn’t have a concrete estimate of when that could happen. Her best guess? “Multiple decades.”
The social robots we have today can complete simple tasks in response to voice commands, and Jibo can even engage in simple conversations. It’s even designed to teach a child a language, or teach an adult a recipe. But in the future, Mitchell reports, machine learning will enable such robots to learn and adapt to their owners, and engage in meaningful conversations with people.
As such, roboticists debate whether robots should take on humanoid characteristics, particularly when those robots are designed to provide companionship or care for the elderly, for isolated individuals, or for people with autism. Matuszek sees value in thinking of robots as companions, and says that “people in general tend to personify their devices whether or not they’re cute.” So if a robot is teaching your kids to play the piano, Mitchell surmises, the humanoid approach could be best. But if you just need a robot to turn off your lights, it doesn’t need to be anthropomorphic.
Robots have a lot of potential to function as smart home hubs in the near future; the primary challenge to overcome is building compatibility with the range of disparate technical standards used by smart home devices. That seems farfetched to early adopters of today’s robots, who find that it’s sometimes difficult to get the technology to understand simple commands. But eventually, a robot will be able to dim your lights on command, or better yet, understand where you are and what you want it to do without your needing to ask.
Researchers are continually developing robots that are more and more adaptable. And as BBC reported recently, that extends not only to their ability to learn from you, but to their ability to cope when accidents and household mishaps happen. A team of researchers recently debuted an algorithm that enables robots to adapt quickly when they are damaged, using a system that quickly filters out the ineffective strategies that the machine could try as it continues to complete the ask to which it’s been set.
That means that when roboticists successfully create more sophisticated, helpful robots than those that are currently available, it may not be entirely incapacitated if it suffers some bumps or spills while going about its household chores. The research could enable robots that are more robust and more adaptable than today’s machines. But, like the rest of the research going on in the field, it also points to the long distance that we have to go until robots will truly be useful for the complex tasks we like to imagine them completing, whether that’s managing preparations for dinner or simply being a companion that you can converse with when you’d otherwise come home to a silent house.