Robot-makers often test the limits of the possible and make ingenious new technology. Can we make this? Can we make it do that twice as fast? Robot users, on the other hand, have different concerns, like “is it simple enough for my kid to understand?” or “if it gets knocked over, will it break?” or “how do I tell it not to come so close?” or “what happens if we move the table to the other side of the room?”
The difference in those worldviews has sometimes led to robots that looked great in the lab but that, in the outside world, make more work for humans rather than less, or that do tasks people don’t need done. Or instances of what I call MORD — the moment of robotic disappointment, when a hyped-up user finds a robot to be much less than she expected.
As robots appear more often in the lives of ordinary people, it becomes ever more important to connect the worlds of makers and users.
That was one of the main subjects of my conversation with Odest Chadwicke Jenkins of the University of Michigan. He wants to make robots that people can understand and really want to use — robots that can clean up a messy room; robots you can show how to vacuum; robots that are designed to do what elderly people want, not what engineers imagine elderly people want. That’s an engineering challenge, but also a psychological and cultural one.
Jenkins and I talked about the reasons humanoid robots are having a moment, why he expects their growth to increase, and about how and engineers need to listen to non-roboticists. We also talked about his own experiences with MORD. Give us a listen!
Caveat: As always, the transcript is AI generated, and thus contains some errors.
Making the Robots People Actually Need