Want your customers to have a positive, forgiving attitude towards your robot employees? Then make your robots human – but not too much! Kai Chi Yam and David De Cremer of the National University of Singapore fill us in on their intriguing work on the subject.
Organisations have started to adopt the use of service robots at an increasing pace. The International Federation of Robotics, for example, found that there was a 61-percent increase in the sales of service robots in 2019. As service robots clearly become more prevalent, the next challenge concerns whether they should be made more humanlike. The New York Times, for example, recently published an article entitled “Should Robots have a Face?”. Anecdotes reported in that article suggest that consumers generally react more positively to robots with faces. In our research, recently accepted by the Journal of Applied Psychology, we tested this hypothesis systemically with two randomised experiments.
We used both psychological and physical means to manipulate how participants thought about the robots in our studies. In Study 1, we used a concept called anthropomorphism – imbuing robots with humanlike characteristics, motivations, intentions and emotions – and explored whether anthropomorphised robots become more likeable. To test this hypothesis first in the field, we went to the world’s first robot-staffed hotel, in Japan, which uses a raptor robot and a humanoid lady as receptionists.
As hotel guests checked into the hotel, we randomly approached half of them and asked them to treat the robots in the hotel as if they had humanlike traits and write a brief description of the anthropomorphised robots, a previously validated manipulation of anthropomorphism. A day later, we approached guests as they checked out and asked them to report on any service failures by the robots that they had encountered during their stay, and their overall satisfaction with the hotel.
We found that the guests we had instructed to think and write about the robots in an anthropomorphised way later reported higher overall satisfaction with the hotel. Notably, we also found that robot service failures were not uncommon. But what is interesting is that guests who were in the anthropomorphism condition were much more forgiving of such failures, compared to guests in the control condition, who were not asked to think of the robots in anthropomorphised ways. This finding is particularly important, given that robots are not yet able to navigate and coordinate well with humans, which means that service failures are likely and common. As a result, companies that today rely on the use of robots in the service sector will definitely need to create the right conditions under which their customers are willing to be tolerant towards these types of failures.
Armed with this insight, we sought to replicate our findings in a more controlled, laboratory setting. In a second experiment, we used a robot to serve participants food and we manipulated anthropomorphism via three factors – voice, name and face. In the anthropomorphism condition, the robot introduced “herself” as “Allison”. “She” also spoke in a normal female voice in an American accent. Finally, the screen mounted on the robot displayed a smiley face during the study and its lips would move while “speaking”. In the control condition, the robot introduced itself as “robotic arm 57174”, spoke in a mechanistic voice, and only displayed a blank screen.
All of our findings were replicated; the participants who interacted with the anthropomorphised robot reported higher satisfaction with the robot. In addition, when the robot failed to deliver the correct food choice, participants in the anthropomorphised condition were much more forgiving of its error.
Given these findings, we suggest that organisations adopting robots in their workforce should do the following to maximise consumer satisfaction:
1. Humanise your robot employees whenever possible. This goes beyond just giving them a face.
Organisations can change the way and the tone with which robots speak, as well as their accent. Our work and that of others have shown that robots who speak with a humanlike voice and a local accent and use slang are more likeable.
2. Making robots more humanised will definitely matter most when automation is used in areas where service failures may still emerge from time to time.
Under those circumstances, your customers’ satisfaction will be shaped largely by how tolerant they want to be towards robotic failures. Ensuring that your robots have human characteristics (such as a face) may then go a long way to ensuring that kind of tolerance.
3. At the same time, however, it is important to avoid making robots look too much like a human.
Indeed, while robots that have a face are generally more likeable, animated faces should be preferred over faces that closely resemble the physical features of humans. Research suggests that we react positively to robots with physical features that are familiar to humans (i.e., animated faces), but push that a little further – close to, but not quite, human – and then they become disturbing, creepy and untrustworthy. This phenomenon is known as the uncanny valley, which refers to the uneasiness people often feel when seeing humanoid robots that resemble actual human beings.
About the Authors
Kai Chi (Sam) Yam is an Associate Professor and Dean’s Chair at the National University of Singapore Business School. Sam is also a research affiliate at AiTH and received his PhD in Organizational Behavior from the University of Washington. Sam’s research focuses primarily on behavioral ethics, leadership, humor, and the future of work. His work has been published in premier journals such as Proceedings of the National Academy of Sciences, Academy of Management Journal, Academy of Management Review, British Medical Journal, Journal of Applied Psychology, Journal of Personality and Social Psychology, Organizational Behavior and Human Decision Processes, and Personnel Psychology. In 2016, Sam was named by Poets and Quants as one of the Best 40 under 40 Business Professors in the world.
David De Cremer is a Provost’s chair and professor in management and organisations at NUS Business School, National University of Singapore. He is the founder and director of the Center on AI Technology for Humankind at NUS Business school. Before moving to NUS, he was the KPMG endowed chaired professor in management studies at Judge Business School, University of Cambridge. He has been named one of the world’s top 30 management gurus and speakers in 2020 by the organisation GlobalGurus and recently published the book “Leadership by algorithm: Who leads and who follows in the AI era?”