Should robots take control? MIT scientists say yes
Artificial intelligence -- some fear it, some welcome it. New research from MIT shows that robots are not only more efficient than humans at controlling human workers, but that the workers prefer receiving orders from robots than from another person in a higher-up position.
People running manufacturing companies reportedly believe that robots could do a good job of supervising human workers, replacing humans doing things like aisle-running. However, they think that human workers might not enjoy working under a robot. We might start to feel worthless, like we could easily be replaced by a robot.
So MIT researchers set out to find a way that robots could help streamline certain processes without trampling the feelings of human workers.
"In our research we were seeking to find that sweet spot for ensuring that the human workforce is both satisfied and productive," says project lead Matthew Gombolay, a PhD candidate at MIT's Computer Science and Artificial Intelligence Lab (CSAIL). "We discovered that the answer is to actually give machines more autonomy, if it helps people to work together more fluently with robot teammates."
The study was comprised of three task groups, each consisting of two humans and one robot. In the first group, all tasks were assigned by a human. In the second, all tasks were assigned by a robot. In the third, one human assigned her own tasks to herself, and the other human was assigned tasks by a robot.
The study found that the fully autonomous group, where the robot assigned tasks to both humans, was the most efficient and effective. More surprising, the human workers said that they preferred the groups where the robots took more control. They were more likely than other groups to say that robots "better understood them."
Gombolay said that this idea could be used to help delegate tasks via an algorithm devised by humans, where the tasks would then be carried out by humans. This research doesn't mean that the robots are ready to take full control yet.
"Instead of coming up with a plan by hand, it's about developing tools to help create plans automatically," Gombolay said.
This could provide an easier way for manufacturers to delegate tasks in the future, rather than having human supervisors do the tedious job of watching over workers.
The team made a video explaining the results of the study. You can watch the video online here.