A lab experiment investigated how college students respond to touch from a humanoid robot during conversation. The findings, published in the journal PLOS One, revealed that students who received touch from the robot (pats on the hand) felt more positive affect during the interaction. Moreover, the students were more likely to comply with a request from the robot if it was accompanied by touch.
Interpersonal touch — from human to human — has been linked to numerous benefits such as reduced stress and improved immune functioning. Some studies have suggested that touch from a robot can also elicit positive reactions in humans, suggesting the potential for a new form of therapy called “robotherapy.” But findings from these studies have been largely mixed and the studies themselves were not without limitations.
Study authors Laura Hoffmann and Nicole C. Krämer note that previous studies have almost exclusively evaluated mutual touch between robots and humans. The researchers were interested in isolating the effects of touch initiated from a robot to a human. To do this, they designed a lab experiment where participants engaged in a one-on-one interaction with a robot who either patted their hand during conversation or not.
A total of 48 students from a European university took part in the lab study. The students were told they would be having a conversation with a robot counselor and were randomly assigned to a touch or no-touch condition. In both conditions, the participants were filmed as they engaged in a conversation with a humanoid robot that was being controlled by an experimenter in a separate room. The robot was Softbank Robotics’ NAO robot and was roughly two feet tall with a plastic body that included eyes, a mouth, and hands.
In the touch condition, the robot patted the back of the participant’s left hand on four separate occasions during the conversation. One of these occasions was accompanied by a request from the robot — asking the student whether or not they were interested in taking a particular business course. During the no-touch condition, the robot did not touch the participant during the conversation. At the end of the interaction, participants completed a questionnaire that assessed their feelings about the experience.
When the researchers analyzed the video recordings to evaluate participants’ reactions to being touched, it appeared that the students did not react negatively to the robot’s touch — no students withdrew their hand from the robot, although two students did raise their eyebrows the first time the robot touched them. Smiling and laughing were common reactions, occurring after more than half of all touch instances. Moreover, laughing was significantly more common in the touch compared to the no-touch condition.
Interestingly, participants in the touch condition were also significantly more likely to comply with the robot’s request, saying they would consider taking the suggested business course. Hoffmann and Krämer note that when a person touches someone while asking for help, it tends to be perceived as an indication of trust in the person they are touching. It could be that people who received the touch from the robot were unconsciously taking this as a sign of trust and thus more willing to comply with the robot’s request. If robot touch can indeed increase compliance, the researchers say that touch from a robot can potentially be used to encourage people to engage in healthy behaviors like taking their medication or exercising.
Participants’ ratings of the robot’s touch were overall positive, with judgments leaning toward “functional, warm, positive, appropriate, pleasant, natural, weak and not painful.” The students in the touch condition also reported feeling more positive affect during the counseling interaction compared to those in the no-touch condition and less negative affect following the interaction.
The study authors say it is unclear how much of this positive experience was down to the robot’s touch behavior and how much was down to the “overall friendly appearance” of the small, plastic robot. They suggest that a similar experiment with a more mechanical-looking robot might not yield the same results. Further study is necessary, the authors say, noting that “more research is needed to understand what the exact differences between robot touch and human touch actually are.“
The study, “The persuasive power of robot touch. Behavioral and evaluative consequences of non-functional touch from a robot”, was authored by Laura Hoffmann and Nicole C. Krämer.
