Robots are increasingly being deployed for services with critical social components, including caring for children and the elderly, education, entertainment, companionship, and so on. However, once the robots are available to the public, the day-to-day interactions that these robots have with users are often highly dissimilar to robot interaction studies conducted within the confines of laboratories. Attempting to conduct human-robot interaction (HRI) studies in a real-world environment, however, often produces results that do not easily lend themselves to the same scientific standards that the literature has come to expect, which are easily produced in laboratory settings. The goal of this workshop is to bring together people who are already deploying social robots in such non-laboratory scenarios, to share methods used and challenges encountered along the way.
Traditional psychology studies hold that experimental setup and data collection must be highly controlled, in order to make it possible to isolate different variables and reach reliable conclusions with reproduceable methodology. However, the reverse side of these high standards is the removal of such studies from real-world scenarios. In the field of HRI in particular, it is highly desirable to incorporate real-world scenarios into a study, as the variables that would be controlled or removed in a laboratory setting may be the very same factors that lead to unexpected reactions and responses.
Social robots and artificial intelligence (AI) differ from other technological tools in our lives, because there is necessarily an emotional and social aspect to the interaction uniquely built into the design. There are robots being developed for companionship, health care, education, physical and emotional support, and other such roles. Whether these robots and AIs are designed for a single specialized function or a variety of different functions, their interaction with humans is necessary. The more numerous the functions the robot can perform, the more complicated the interaction with the human is likely to become. This goes beyond mere functionality, as the human users must be capable of trusting the robot or AI, and to communicate with it.
Thus technologists are faced with the challenge of designing robots that interact intuitively with the user and also fulfill the role for which they were designed.
In technology, there can be a tendency to place the onus of learning new technology on the user. However, this necessarily means that the user cannot form any intuitive, natural connection to the technology—a connection which, in particular in social contexts, would be highly beneficial. Demanding that the users adapt to an interface that does not account for our natural inclinations is short-sighted in the long term. Yet we don't know enough about social intelligence and which aspects create attachment to adequately design, for example, an educational robot capable of engaging students by simulating charisma, interest and passion.
Moreover, as social robots are becoming more commonplace, and people are starting to make emotional connections with them, we need to study such human-robots relations carefully and understand what these attachments mean before letting robots into the wild. One goal of our seminar is to focus on the emotional aspects of the human-robot relations and to explore what new roles we can give to robots in the future: study desirable and less desirable future scenarios in order to craft the preferable future.
Our goal is to identify usability and design issues for the future social robots and the scientific and technological barriers towards successful innovation, with focus on studies conducted in real-world settings, outside of the laboratory. By the end of the workshop, we hope to generate a foundation for ethical and legal guidelines for deploying such robots in various social settings.