If robots aren’t everywhere yet, they’re well on their way. Robots are getting better vision, dexterity, and intelligence, so they’ll be used to do more things in more settings. And they won’t be confined to special rooms or behind barriers on the factory floor, either; they’ll be working right next to us. Consider that robots are already used for some pretty personal tasks such as:
If robots aren’t quite colleagues, they’re already more than tools
So, if we can expect to encounter robots in our lives, what do we want from the experience? Engineers and psychologists involved with human-robot interaction are working on that question and they’ve come up with some early answers. Their results will help define how comfortable we are in a world of shared human and machine intelligence. Two topics of current research include appearance and behavior.
If it looks like a human . . .
A new study shows that appearances matter when it comes to robots. After all, would anyone feel comfortable working alongside an office Terminator? We tend to prefer robots that look like us, but the feeling is nuanced. Experimenters at Georgia
Tech developed three types of “faces” for possible robot designs: human, robotic (machine-like), and a hybrid of the two. They then showed these faces to two groups of people divided by age. The younger group preferred the robot faces while the older group preferred the human faces. Almost no one preferred the hybrids.
The age difference all but disappeared, however, when people were asked again in the context of the robot’s purpose, i.e., the task it was designed for. Most preferred a human face for robots that perform social tasks like teaching or providing information (i.e., a robot banker, while robots with machine-like faces were more acceptable for things like household chores (good news for the Roomba!). Apparently, people tended to view robots as smarter if they had human-like faces. There’s a downside to looking like us, though; if the robot task was very personal, like bathing assistance, some people were uncomfortable if the face was too human.
And acts like a human . . .
It only takes a few autonomous behaviors to evoke a connection between people and machines. In other words, if we see an example of independent action from machine, we tend to attribute some level of purpose and motivation to it. This trait is what involves us emotionally with movie animations and computer games – behaviors without physical reality. Similarly, robot behaviors are powerful triggers to human responses. Japanese researchers may be doing some of the most advanced work in this area, where natural appearance, natural voice, and natural social behaviors are all being pursued in an effort to achieve almost android-like effects. Human connection, however, can be generated with far less: Even the limited cues offered by the Japanese Paro robot, used as a therapeutic companion, generate feelings of empathy among patients that interact with it.
The trick is to clearly define what a robot purports to be. While even limited actions can evoke feelings of empathy with another behaving entity (even an artificial one), the effect can be broken when claimed behaviors don’t meet expectations (like the disappointment of many iPhone users regarding Siri’s capabilities).
We respond in kind
The strongest examples of human connection with robots come from the battlefield, where soldiers grow attached to the bomb disposal machines that keep them safe. Bomb disposal is a life-and-death task, of course, so feelings are bound to run high. Nevertheless, it’s not uncommon for soldiers to give their robots nicknames and to feel anguish and loss if a robot is damaged during operations. One research study even revealed instances of soldiers giving mock awards to their robots after successful missions and mock funerals when the machines were destroyed.
The 2012 film Robot and Frank used a different basis for a human-robot relationship but illustrated many of the same response patterns found among soldiers. There don’t seem to be any fundamental barriers, therefore, to building genuine relationships with intelligent machines.
Increased capabilities mean increased roles for robots. Although this will almost certainly yield better labor productivity, safety, and quality of life, the alarm over lost job opportunities has already been sounded. Proliferating use of robots might not be entirely good news for everyone.
Hopefully, the larger process of studying human-robot interactions will also reveal solutions to issues of balancing people and machines to best complete society’s tasks. As the physical and behavioral features of robots are being examined in human terms, perhaps their roles will be examined in human terms, as well.