Researchers test how much people really trust robots
A new study from the Georgia Institute of Technology tested human response to robots in emergency situations, and had some interesting results.
Titled “Emergency Guide Robot,” the study found that during staged building fires, human test subjects would frequently follow robot instructions. But that’s not the interesting part — testing discovered that the subjects would follow robot instructions even after that robot had already proven itself to be unreliable.
“People seem to believe that these robotic systems know more about the world than they really do, and that they would never make mistakes or have any kind of fault. In our studies, test subjects followed the robot’s directions even to the point where it might have put them in danger had this been a real emergency.” — Alan Wagner, a senior research engineer in the Georgia Tech Research Institute (GTRI).
Blind trust?
Researchers analyzed the behavior of 42 participants, who were instructed to follow the robot into a conference room where they were to “complete a survey about robots.”
With the true nature of the study masked to the participants, the robot would sometimes erratically lead them to the conference room. For example, the robot would lead participants to the wrong room, or malfunction in other ways. Sometimes, participants were even told that the robot had broken down.
Once the staged building fire took effect, researchers surprisingly watched test subjects follow the robot’s instructions to exit the building safely — no matter how untrustworthy the robot had been leading them to the conference room.
Wagner said that one of the biggest takeaways from the study would be to investigate how to prevent people from trusting robots too much in emergency situations.
In emergencies, humans are too trusting of robots Georgia Tech research finds. | https://t.co/kceUt6dAAn pic.twitter.com/d5xaiC4C6r
— Georgia Tech (@GeorgiaTech) March 2, 2016
First human-robot study of its kind
“Emergency Guide Robot,” which was sponsored in part by the Air Force Office of Scientific Research (AFOSR), is thought to be the first-ever study to analyze human-robot trust in an emergency situation.
Comments are closed.