U. of Michigan study: Humans not tolerant of robot mistakes

As robots become more pervasive in a variety of industries, a new study finds that their human colleagues may not be forgiving when the robots make an error.

According to a University of Michigan news release, research conducted by a team at the university found that humans begin to distrust robotic counterparts when mistakes are made and regaining trust can be hard. The release noted that the project included 240 participants, working alongside robots, and the machines made mistakes. 

After three errors, none of the efforts by the robot were able to restore faith that the humans originally had in the machines. Connor Esterwood, lead author of the study, noted in the release that the results can have far-reaching implications for robot builders and users.

"Researchers must develop more effective repair strategies to help robots better repair trust after these mistakes," he said in the release. "Also, robots need to be sure that they have mastered a novel task before attempting to repair a human’s trust in them. If not, they risk losing a human’s trust in them in a way that cannot be recovered.”

The strategies used by the robots include denials, apologies, explanations and promises, with some proving effective after the first error, and becoming less effective with subsequent mistakes, the university said. 

Study co-author Lionel Robert pointed out that the findings demonstrate theories of misinforming, forgetting, informing and forgiving. After a trio of errors and fixes, trust wasn’t able to be fully restored, which could cross over to human-human trust issues.

The University of Michigan release noted the results of the study were published in the journal Computers in Human Behavior and demonstrate how important it is to ensure robots are reliable and can perform their tasks without frequent errors. The study shows how important it is to have an effective repair strategy to retain trust in a robot when a mistake is made.