There Is No Point Making Robots Look and Act Like Humans
Posted in: R&D and Inventions, Robots, Today's Chili
By Olivia Solon, Wired UK
The Terminator, C-3PO, the Cylons, and the Jetsons’ robotic maid Rosie are all highly agile and memorable humanoid robots from science fiction. They are intelligent, nimble, dexterous, autonomous and you never see them plugged into an energy source, waiting to refuel.
Now take Asimo, described by its maker Honda as “the world’s most advanced humanoid robot.”
There is no denying that the robot is spectacular, walking and even running with ease on two legs, responding to voice commands and mapping its environment using camera “eyes”. However, it requires at least one person (and preferably two) to control it, almost a day to set up and can operate for just one hour on a single 51.8v lithium ion battery which requires three hours to recharge.
Despite wonderful technological advances, the humanoids that we have created so far fall extremely short of those we have imagined within movies and books. At the Innorobo robotics summit in Lyon last week, I saw all sorts of delightful and adorable humanoids (Acroban, Nao and Darwin to name a few), but witnessed them fall over, run out of batteries, fail to understand people and break down completely. These are lovely, and very expensive, toys — as opposed to viable butlers, nurses or companions. Our expectations are, frankly, too great.
We need to shift away from the idea of a humanoid butler who comes to the house, understands our needs, and uses our hoover, our washing machine and our oven and think about how to “robotise” the things that we already have in our house. This is not only representative of the way things are moving (particularly when it comes to the internet of things) but is also likely to be much more cost-effective.
Francesco Mondada, researcher in AI and robotics at Switzerland’s national robotics centre, agrees. He told the Innorobot audience: “We should improve objects instead of creating one device that is exterior to the other objects that can interact with the regular household. Instead of having a robot butler to park my car, we should be getting the car to park itself. This is the way things are moving.”
He added: “The day that a humanoid understands all of our feelings even when my wife doesn’t, will be great. But not even someone I’ve been married to for 25 years can understand me, so a robot has no chance.”
Humanoid robots might be wonderful to behold, but the robots that are gaining traction in the real world resemble other things — such as the disc-shaped Roomba, the military Packbot and the robotic seal companion Paro.
Roboticists are keen to create carer robots to help tend to the elderly — with nursing shortages and an ageing population, there could finally be a viable market for domestic robots. However, there is a big debate as to whether these devices need to be purely functional or whether they need to be social.
Those on the functional side of the debate say that elderly people need help performing physical tasks, remembering to take their medicine and generally making daily activities easier. They do not need a robotic “friend”. Mondada argued that a person’s social life will of course be affected if they are physically dependent, but this doesn’t mean we should be creating social, humanoid robots for them to talk to because they can’t go and see their friends.
They need a functional robot to help them to do the things that they can no longer do. The most social interaction they want from a robot is the level of engagement one might otherwise get from a dog or a cat — it should not be a substitute for visits from the family.
Those on the social side say that in order to garner acceptance from technophobes, robots need to be appealing and social. They need to be able to communicate in the same way a human does, rather than rely on complicated interfaces and provide companionship where necessary.
The second argument misses the point. It seems entirely possible that you can have a simple communctions interface for a purely functional robot. You could summon a telepresence device to call your friend or relative using voice activation, without needing the telepresence device to have its own personality. A domestic robot could learn about your preferences and habits without needing an opinion on the latest news.
Of course, in the future we will be able to overcome the shortcomings of current humanoids and create jaw-droppingly agile bipedal, emotionally responsive devices that will know you better than your wife does.
But beyond the realms of research, what’s the point?
