Humans May Value Relatability Over Efficiency in Robots


It's okay, BERT. We forgive you. Image credit: University of Bristol

The most important feature in a robot that people will want to use may be the ability to apologize. That’s according to a new study from University College London and the University of Bristol [PDF] that examined how a robot’s expressiveness can influence a user's experience. As Mashable reports, participants were more willing to forgive errors made by a machine if it expressed regret.

For their experiment, researchers had three different versions of the same robot, BERT2, bring subjects the ingredients to make an omelet. BERT A completed the task without error, BERT B dropped an egg at some point, and BERT C dropped an egg while also acknowledging the error with regret. When he made the mistake, BERT C said “I’m sorry” and conveyed a look of disappointment. He then fixed the mistake by approaching the task from a different angle.

Even though BERT C wasn’t the most efficient bot of the bunch, he won the favor of participants. He was perceived to be the fastest at completing the task, while in reality his communication slowed him down by 50 percent. Fifteen out of 21 volunteers said they’d choose BERT C for work-related tasks, but not necessarily because they believed he was the best robot for the job. When asked whether they would hire BERT C as a kitchen assistant, many people were reluctant to give an answer. One participant imagined the robot looking hurt when he responded “no” and another participant lied to spare the machine’s feelings.

The study, the researchers write, "suggests that a robot effectively demonstrating apparent emotions, such as regret and enthusiasm, and awareness of its error, influences the user experience in such a way that dissatisfaction with its erroneous behavior is significantly tempered, if not forgiven, with a corresponding effect on trust."

The data is limited in scope and has yet to be peer-reviewed, but previous research supports the idea of people connecting to robots on an emotional level. A study published in 2015 found that people empathize with robots they perceive to be in pain (despite their lack of nerve endings). With more data to back up this concept, guilt-tripping capabilities could come standard in robots of the future.

[h/t Mashable]

Know of something you think we should cover? Email us at