top of page


議論先導者:Dr. Soheil Keshmiri(ATRリサーチャー)

Dr. Keshmiri is a researcher in Hiroshi Ishiguro Laboratories (HIL) at Advanced Telecommunications Research Institute International (ATR). His research focuses on modeling of the individuals’ neurophysiological responses during their interaction with human-like agents. He is a recipient of the ATR Research and Development Award for Excellent Research in 2017 and 2019.​

“Real” versus “Artificial” : How “Uncanny” is the “Uncanny Valley Hypothesis”

As the field of robotics and artificial intelligence mature, the prospect of synthetic intelligent agents that will live alongside the humans becomes less of a science fiction and more of a near-future real-life event. Although it is true that the social and intelligent capabilities of these agents are currently limited, the fast pace of scientific and technological advances will surely overcome this limitation, making it a thing of the past. However, such advances do not necessarily warrant a change in our perspective on so called “real” versus “artificial” distinction. After all, we are presumably the “real being” who are capable of making “tools” to serve our purposes and leave them in the attic of history once we have no further use for them. If so, how will we take these new fast emerging intelligent entities? Will we treat them as “tools”, unplugging them as it pleases us? Or will we take them as “companions” and “partners” sharing our planet with? Will we trust them? Will we be able to get close to them and feel comfortable around them? Will we ever fall in love

with them?


Some argue that the Uncanny-Valley (UV) hypothesis has already addressed this question. That humans prefer anthropomorphic (human-like) agents but reject them if they become too human-like. The immediate implication of this viewpoint is that the more intelligent and realistic these synthetic agents become, the less trustworthy we will find them, thereby drawing a dire prospect on any potential symbiotic life between “us” the “real” and them the “artificial.” I personally think that this viewpoint does not give us the full picture.


In this talk, we will make an attempt to better understand the “uncanny valley effect” by finding its analogies and examples in different aspects of our daily life in which no robot exists. This will allow us to dissociate this concept from synthetic agents and appreciate its presence in the way we make sense of the world around us to increase our ability to survive and thrive. While doing so, we will discuss the role of our brain in weighing our experiences. This will lead us to realization that there is nothing particularly “uncanny” about the synthetic agents: The entire world can be “uncanny” if it does not fit into our expectations!


By the end of this talk, we will hopefully have a better understanding of what is “uncanny” about “uncanny valley hypothesis” and feel more confident to take such questions as “will we feel uncomfortable if one day a robot shows up at our doorstep and says: Hi, do you wanna go watch a movie tonight?!” I know my answer already: Well, I think we might if they don’t act like we all do and we won’t if they do. I think “the uncanny” part of the problem is because today’s robots don’t fit in our brain very well and by that I mean they act and talk and do things very differently from the way we do.

bottom of page