Interactions with Artificial Humans

A main assumption in the project Interactions with Artificial Humans is that studies of how we humans react when we engage with non-human agents would be useful to facilitate our lives when such agents become more common.

But such studies can be useful for other purposes, too. One important by-product of studying interactions with non-humans is that we may actually learn more about us humans also in a human-to-human context. The main reason is that many studies of humans’ interactions with non-humans, including several studies in the project Interactions with Artificial Humans, are based on an anthropomorphism assumptions (i.e., we tend to react to humanlike non-humans in ways that are similar to how we react to real humans). Given anthropomorphism, it becomes relevant to use theories and findings from human-to-human interactions for studies of humans’ interaction with non-humans – and this is exactly what researchers do in many studies. For example, if you are interested in how humans react to a non-human that displays happiness, it becomes useful to use theories about how humans react to other humans who display happiness.

Our theories of human behavior in human-to-human settings, however, are far from complete. Indeed, there is so much more to learn. This, then, is the reason why there is still a strong need for developing theories in areas such as social psychology and consumer behavior. And a somewhat peculiar aspect of doing research is that theory improvement is particularly likely to occur when an existing theory is used in a study with empirical data. In such studies, when the researcher is trying to make sense of his or her data in the light of a theory, it often becomes clear that a theory is not able to fully deal with the data. The very activity of testing a theory, to see if it survives the encounter with data, is thus one important way in which theory can be improved.

So, when we apply theories about human-to-human interactions in settings in which humans interact with non-humans, we increase theories’ exposure to data. And thereby we also increase the opportunities to identify limitations, flaws and weaknesses of theories, so that we get a chance to improve them.