Högskolan i Skövde

The development of robots that are able to interact socially with humans is still in its early stages. Since the beginning of the 2000's, these (so called) social robots have started to emerge in a variety of settings. Along with the emergence of social robots, there has been a parallel interest to study the different aspects of having humans interact with robots socially. There are several motivations behind developing and studying social robots; social robots may be used as test beds to study human behavior, as tools for humans to achieve certain tasks in specific contexts, or as interaction partners and thus viewed as social agents. These three perspectives often draw on the assumption that human-robot interaction (HRI) is similar to human-human interaction. Thus, humans tend to expect human-like abilities in social robots, often mismatching the robots' actual capabilities.

In this thesis proposal, expectations of social robots are the focal point. Expectations are, in any aspect of life and not just in HRI, underlying and ever present mechanisms of human behavior. Expectations are defined as believed probabilities of future events that set the stage for the human belief system which guides our behavior, hopes, and intentions. Expectations are based on direct experience, other people, and beliefs. Once an expectation is set, it is accompanied by either positive or negative affect which turns to behavior and performance. Thus, expectations are crucial in human behavior, including when interacting with social robots. What makes social robots more rare than other technical artifacts such as computers, is the lack of personal experience for many humans. High expectations happen especially with social robots as they are purposely designed to look and behave like humans, thus creating ethical implications as it can be considered deceptive design.  Expectations are therefore usually built on beliefs based on the portrayal of social robots in media. When humans interact with social robots, they will usually have high expectations which ultimately has an effect on how successful the interaction will be. This creates a gap between what is expected, and what the robots are actually capable of.

Expectations are thus an underlying factor in interaction with any artifact, and there is a need to get a deeper understanding of how these expectations affect HRI. Once we have gained a richer understanding of how expectations affect HRI, we can narrow the expectation gap, and create more successful interactions between humans and social robots in society. With this in mind, the aim of my PhD work is to investigate the role expectations play when interacting socially with robots, including the subsequent ethical implications of such expectations. My four objectives are to (1) theoretically identify existing research on expectations in HRI, (2) empirically investigate expectations in HRI, (3) synthesize the obtained findings from objective 1 and 2 to create an interdisciplinary theoretical framework of expectations in HRI, and (4) address the ethical implications of expectations in HRI. In this thesis proposal, I present what I have done so far to reach these objectives, as well as my research plan moving forward towards my dissertation. The intended contributions of my PhD work is to theoretically and empirically characterize the role and relevance of humans' expectations when interacting with social robots with the goal to narrow the social robot expectation gap.