Workers Want Robots In Charge

By Joelle Renstrom | Updated

CSAIL
Photo credit: CSAIL

That’s right. For all our fears of succumbing to robotic overlords, it turns out that we want robots to hold dominion over us — at least, when it comes to the workplace.

Even though many people believe robots and other automated systems will put many out of work (others believe they will usher in a new era of innovation and resourcefulness), research conducted by MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) reveals that workers actually prefer for robots to take the lead in manufacturing tasks. The study explores the two sides of robotic workers: on the upside, they free humans from tasks characterized by the “three D’s” — tasks that are dirty, dangerous, and/or dull. Of course, if robots do assume those jobs, what’s left for humans? Oversight? Programming? Perhaps collaboration with the robots?

It’s still possible, if not likely, that human workers will be upset about being replaced, and that some will feel almost fatally dispensable. Is there a middle ground? CSAIL’s study tried to look for what Matthew Gombolay, project lead, calls the “sweet spot” that makes for a productive workplace, as well as satisfied employees. According to the research, the answer is robotic autonomy, which makes it possible for humans and robots to truly work together.

The study assigned groups of two humans and one robot to work together in three different configurations. The first was manual — i.e., humans assigned all tasks. The second was automatic, with robots assigning all tasks. The third was somewhere in the middle, semi-autonomous, meaning that one of the people in the group assigned his own tasks, while the robot assigned tasks to the other person. The team then measured the productivity of the groups, as well as which strategy the human workers preferred.

Turns out, the automatic scenario in which the robot delegates tasks was the most productive, and the one human workers preferred. In addition to increasing efficiency, human workers had the sense that robots in charge “better understood them.” It’s also worth noting that the robots in charge aren’t sentient beings; they run on algorithms designed by humans. Those algorithms are flexible and able to accommodate problems or on-the-fly developments, but still, they form the “brain” of the bots. In addition to the bots themselves, such algorithms could be helpful just about anywhere — in hospitals, battlefields, or disaster areas, to name a few. One could also argue that they’re preparing us for the future, when we may be receiving tasks from robots whose programming we don’t entirely control.