An empirical analysis of the uncanny valley of human-like robots through a choice experiment in Korea

As humanoid robots and artificial intelligence technologies advance, people’s concerns about non-human entities becoming like or replacing humans are increasing. However, considering the prospect of coexistence with robots in the near future, helping people naturally adapt to this new relationship has become important. We estimated people’s utility function over social robots’ appearance, decision-making ability, voice, naturalness of motion, and emotion perception and reaction through a choice experiment. The results revealed the existence of a saddle point in people’s utility—known as the uncanny valley—as social robots’ appearance becomes more human-like. Regarding decision-making ability, people’s preference decreased as the decision autonomy of human-faced social robots increased. Our findings indicate that simply mimicking humans is insufficient for social robots to be embraced as new members of society, highlighting the need for more comprehensive considerations in their development.