Researchers at MIT’s Computer Science and Artificial Intelligence Lab just revealed a robot they’ve programmed to give human coworkers insightful suggestions to increase operational efficiency. They’ve recently published papers in which they’ve put their robot to work in different settings. Most notably, teams successfully demonstrated how the robot, by observing humans work at Beth Israel Deaconess Medical Center, learn how to help nurses schedule tasks.

As seen in the video, the programmed NAO humanoid robot learned the ins and outs of the labor ward at Beth Israel. It was thereby able to make operational recommendations, such as where to move patients and how to assign nurses to procedures like C-sections.

“The aim of the work was to develop artificial intelligence that can learn from people about how the labor and delivery unit works, so that robots can better anticipate how to be helpful or when to stay out of the way – and maybe even help by collaborating in making challenging decisions,” Julie Shah, MIT Professor and senior author on the research paper, explained in a statement.

Coordination in Beth Israel’s labor ward is an intricate dance. According to MIT, the head nurse there simultaneously has to coordinate 10 nurses, 20 patients and 20 rooms, bringing the number of possible scheduling options to two to the one millionth power.

The researchers developed a scheduling policy allowing the robot to learn the best possible scheduling options – and the worst – and to respond appropriately to new situations it hasn’t seen within the ward.

You can put the robot on the labor floor, and… it will understand how to coordinate an efficient schedule.

“Rather than considering actions in isolation of each other, we crafted a model that understands why one action is better than the alternatives,” Shah shared. “By considering all such comparisons, you can learn to recommend which action will be most helpful.”

The policy prevents medical personnel from sitting down and personally ranking every scheduling possibility based on operationally efficient it would be.

“You can put the robot on the labor floor, and, just by watching humans doing the different tasks, it will understand how to coordinate an efficient schedule,” Matthew Gombolay, a PhD candidate at MIT who co-wrote the paper, said.

With the system-enable NAO robot, nurses at Beth Israel accepted the robot’s recommendations 90 percent of the time. They also asked the robot for bad advice and rejected those suggestions at the same rate of 90 percent, demonstrating that the system knows the difference between good and bad recommendations.

The researchers maintain the robot can be adapted for work environments in different industries. They have already been testing to see how it fares in a military application.