Don Norman: Designing For People

Nielsen Norman Group

Robots in the Home: What Might They Do?

Column written for Interactions magazine.

Robots are coming, but what does this mean to ordinary folks? First of all, don't believe all the hype. Lots of hobbyists and small ventures would have you believe that robots are already here, capable of a wide variety of interactions, including health care and monitoring medication compliance, security monitoring, education, errands, and entertainment. Robots are, of course, used in manufacturing, in search and rescue missions, and in the military. But when we get away from industry and the military and discuss machines that are reasonably priced, most of these so-called applications are more imagination than reality, with unreliable mechanisms barely able to get through demonstrations. For everyday home applications, the use of robots is restricted to entertainment, vacuum cleaners, and lawn mowers. Note, however, that the definition of "robot" varies widely, often being used for anything mobile, even though controlled by a human. Personally, I would classify intelligent home appliances as robots: my coffee maker, microwave oven, dishwasher, and clothes washer and dryer have more intelligence and actuators than robot vacuum cleaners — and they are also a lot more expensive. But they don't move around the room, which for many people disqualifies them from the label of "robot."

Given that any successful product for the home must be affordable, reliable, safe, and usable by everyday people, what might a home robot do? And what would it look like? In the home, form probably will follow function. A kitchen robot might be built into the counter space, with dishwasher, pantry, coffee maker, and cooking units all arranged so that they can communicate with one another and pass items readily back and forth. An entertainment robot might take on a humanoid appearance (as in Wow Wee's Robosapien), or animal-like (as in Sony's Aibo). And robots that vacuum or mow lawns will look like, well, vacuum cleaners and lawn mowers.

Making robots work well is incredibly difficult. Their sensory apparatus is limited because sensors are expensive and interpretation (especially common-sense knowledge) is still more suited for research than deployment. Robotic arms are expensive to build and not very reliable. This limits the range of possibilities: Mowing and vacuuming? Sure. Sorting laundry? Hard, but doable. Picking up dirty items around the home? Doubtful. How about assistants for the elderly or those who need medical supervision? This is a booming area of exploration, but I am skeptical. Today's devices are not reliable, versatile, or intelligent enough — not yet, anyway. Moreover, the social aspects of the interaction are far more complex than the technical ones, something the technology-driven enthusiasts typically fail to recognize.

Three likely directions for the future are entertainment, home appliances, and education. We can start with today's existing devices and slowly add on intelligence, manipulative ability, and function. Start small and build. The market for robots that entertain by being cute and cuddly is already well established. The second generation of vacuum cleaners is smarter than the first. Sony's dog gets smarter and less expensive with each new version. We don't yet think of washing machines, microwave ovens, and coffee makers as robots, but why not? They don't move around the house, but they are getting better and smarter every year. And when the coffee maker is connected to the pantry and dishwasher, that will be a home robot worthy of the name: same for the coupling of sorting, washing, drying, and storing clothes.

Education is a powerful possibility. There is already a solid basis of educational devices that aid learning. Today's robots can read aloud in engaging voices. They can be cute and lovable — witness the responses to the multiple quasi-intelligent animals on the toy market. A robot could very well interact with a child, offering educational benefits as well. Why not have the robot help the child learn the alphabet, teach reading, vocabulary, pronunciation, basic arithmetic, maybe basic reasoning? Why not music and art, geography and history? And why restrict it to children? Adults can be willing and active learners.

Now this is a direction worthy of exploration; Robot as teacher. Not to replace school, not to replace human contact and interaction, but to supplement them. The beauty here is that these tasks are well within the abilities of today's devices. They don't require much mobility nor sophisticated manipulators. Many technologists dream of implementing Neil Stephenson's children's tutor in his novel The Diamond Age: Or, a Young Lady's Illustrated Primer. Why not? Here is a worthy challenge.

Don Norman wears many hats, including co-founder of the Nielsen Norman group, Professor at Northwestern University, and author, his latest book being Emotional Design. He lives at Disclaimer for this article: Norman serves on the scientific advisory board of Evolution Robotics, a startup company whose software is used within several successful home robots.