Robots with personality seem more trustworthy
Robots are showing up in all kinds of jobs these days, from manufacturing plants to offices. So, when people suddenly encounter a robot co-worker on the job, will they like it more if it looks like a “him” or “her” and if it cheers for the same football team?
These are some of questions examined in a recent study, “Human-Robot Similarity and Willingness to Work with a Robotic Co-worker,” to be presented at the 13th Annual ACM/IEEE International Conference on Human Robot Interaction in Chicago, Illinois March 5-8. The co-authors are Sangseok You (PhD ’17), a postdoctoral research fellow at Syracuse University, and UMSI associate professor Lionel Robert.
The study comes as more and more robots are replacing or working alongside humans. They are being employed increasingly on assembly lines, in order fulfillment centers and product inspections service centers.
For example, the study notes that Amazon is adding 15,000 robots every year to work alongside human employees in its 20 fulfillment centers, and that robots are expected to replace as much as half of the workforce in the next 10 to 20 years.
As a result, organizations must figure out how to integrate human and robot co-workers. The increasing replacement of robots for human workers in blue collar as well as white collar jobs has caused resentment and distrust of robot co-workers.
“I think people are becoming more apprehensive about this, because they see job losses,” Robert says. “Part of the issue is humanization. When a robot is just a big arm on the assembly line, that’s different.”
The researchers conducted an experimental study with 200 participants on Amazon Mechanical Turk. The participants were randomly assigned one of eight conditions employing surface-level human-robot similarity (physical likeness); and deep-level human-robot similarity (personality, i.e., same vs. different work styles, similar values, attitudes, beliefs and knowledge).
Participants also were asked about how the risk of physical danger – such as for work in the military or on space missions -- affected their willingness to accept robot co-workers.
Making the robots more human in appearance and in interaction via perceived personality traits increased the level of study participants’ trust levels in robot co-workers, Robert said.
The idea of a robot having a “personality” may strike many as odd. But Robert emphasizes many people may be unfamiliar with the huge advances in artificial intelligence (AI) that enable this kind of complex programming in robots.
“There are massive AI systems in place” that perform all kinds of tasks. “People don’t know there are BOTS making decisions behind closed doors, writing articles, even offering therapy.”
So robots can be programmed to be specifically like a human co-worker, Robert says. “Imagine a robot that is a fan of the Philadelphia Eagles,” the football team playing against the New England Patriots in the 2018 Super Bowl. “It could be programmed to be biased toward the Eagles and change your perception of robots to be more human. All this kind of content is built in.”
As with most new technology – autonomous vehicles being the most recent example -- some people are increasingly comfortable with the idea of robot co-workers, while many others are reluctant to accept it.
For instance, Robert says, “I know some people who refused to use a self-check-out at the market because they thought the check-outs were taking someone’s job away. There will be people who accept robots as entities with full rights and privileges. Others will say, ‘No, they shouldn’t be here.’”
Meanwhile, people can expect more robots in their daily work – and daily life, he adds. “I think McDonalds will be using robots instead of human workers. You’ll probably start seeing stores doing this in the next five or six years.”
The paper is available online and will be published in the proceedings of the 13th Annual ACM/IEEE International Conference on Human Robot Interaction (HRI 2018) in March 2018.