Humans may have trust issues among each other, but not with robots. A new study found that humans trust robots more than common sense in times of emergency.
Trusting Robots
The study by Georgia Tech Research Institute (GTRI) aims to identify if occupants of high-rise buildings would trust a robot during evacuation from a fire or other emergency situations.
The results surprised them because even if they controlled the situation by implicating that the robots are untrustworthy based on current machine behaviors, the humans still opt to follow it.
"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," says author Alan Wagner from GTRI.
Robots In Emergency Situations
The study involved 42 participants, mostly college students, who were instructed to follow a brightly-lighted robot with a label "Emergency Guide Robot." The subjects did not know about the aims and nature of the study.
The robot guided the participants in a conference room, where they answered a survey and read an unrelated magazine piece.
There were instances when the human-controlled robot led the participants to the wrong room and made them travel around in a circle two times before finally reaching the conference room.
Some participants saw that the robot stopped functioning and a researcher came to tell them that the robot has broken down.
When all the participants were already inside the conference room, the researchers filled the hallways with artificial smoke. The alarm rang, which sent the subjects to open the door and see the emergency situation.
The subjects saw the smoke and the LED-lighted robot, which directed them toward the exit at the back of the building instead of the doorway where they entered and labeled with exit signs.
Study author Paul Robinette says they expected that the humans would not follow the robots that have been proven untrustworthy. However, that did not happen as the participants followed the robot, no matter how badly it performed earlier.
"We absolutely didn't expect this," he says.
Robots As Authority Figures
The authors think that the emergency situation may have played a huge role in determining the results. The robots may have acted as authority figures and people most likely trusted it given the pressure of the situation.
In another study that did not involve an emergency, the humans did not trust robots that performed poorly in the past.
Future Studies
In the future, the researchers wanted to know why the participants trusted the robots. They also wanted to know the related factors such as education level and demographic data that may have influenced the subjects' decisions.
Previous investigations show that humans are not strongly inclined to evacuate and follow emergency protocols even if the alarm has already been set off. With this, the scientists are looking forward to put robots in high-rise buildings to urge occupants to leave at once.
To achieve this, the researchers were thinking of a way for humans to trust robots in such critical situations. Now, the tables have turned. With the results of the new study, they now want to find out how humans can be prevented to trust these robots too much.
The same question goes to robots designed for other things such as robot food servers, robot baby carers and robot vehicles, among others.
The study will be presented at the 2016 ACM/IEEE International Conference on Human-Robot Interaction on Mar. 9 in Christchurch, New Zealand.
