The Uncanny Valley of Mobile Services December 12, 2011Posted by reto wettach in innovation process, methods, mobile.
In 1970 the Japanese researcher Masahiro Mori presented the concept of the Uncanny Valley:
The uncanny valley is applied to human-like robots or to 3-D computer animations of humans.
A lot of research has been done since 1970:
Some are trying to understand the origin of the Uncanny Valley as Karl F. MacDorman (2005) who relates this feeling of estrangement to the “Terror Management Theory”, which is concerning “how human beings manage their fear of personal extinction“.
Others are expanding the original model as Christoph Bartneck et al. , who are doubting that the uncanny valley in its quite simplistic approach really exists and are demanding a more complex discussion of this hypothesis. In his paper “My Robotic Doppelgänger” (2009) he compared a real person with his robotic Doppelgänger and could not confirm Mori’s hypothesis, however, the “participants were able to distinguish between the human stimulus and the android stimuli“, which I do not think was too difficult… 🙂 However, in my understanding of the uncanny valley, this estrangement is happening in the moment where you are not sure whether something is a robot or a human…
(image source) (Note: Hiroshi Ishiguro is a co-author of the paper)
Hiroshi Ishiguro is even suggesting to expand the framework and to take “behavior” and “appearance” into consideration (“Android Science – Toward a New Cross-Interdisciplinary Framework“, 2007):
I am quoting this research because the behavior aspect of the uncanny valley can also be applied to human-like behavior, which is not visible as an robot or computer generated 3-D-person. In many “intelligent” services one could have the feeling that they come from humans, e.g. in the area of location based services:
The sweet spot 1 “include simple, focused applications like local search or sat nav systems. They provide clear benefits at low cost to the user, and tend to be transparent about their limitations or how they make decisions.”
The sweet spot 2 is like a “close friend”: “incredibly context sensitive, but also has very good sense of discretion, appropriateness and knows how to honour privacy.”
At IxDS we developed a tool for prototyping contextual mobile services, which we called CWS (Contextual Web Services). This system is based on the concept of Wizard-of-Oz-Prototyping, which means that the user does not know nor experience that a human is controlling the contextual web service on his mobile. The operator is working with a PC-based dashboard which shows the profile of the test person, his/her location, the history of interactions and an easy-to-use interface to send/push a notification – with or without interactive elements. The user can give direct feedback whether he like a certain service or not:
We also observed an “uncanny valley” when the suggestions became too personal. In the qualitative feedback following a three-week-test-period many test users indicated that they didn’t like too personalized suggestions – they found them “scary”…