Interaction with Avatars, Robots and the Web

I just returned from the UK where I gave a lecture to a group involved in visual evidence. I had talked about a subject that I dealt with in my graduate work including animated reenactments of crime scenes. It was an interesting conference that disclosed some information that was new to me, not the least of which was my research was still relevant to the field of jury perception.
More recently, I read an article on robotics and their use in teaching autistic children because robots  potential to have endless patients and repetition skills that autistic children need. I have always thought that my brothers work in tactile fabrics that measure pressure could aid in child learning disabilities if designed to measure both pressure of mechanical "hugs" as well as measuring a child's pressure of returned affection or physical interaction as some reference to their satisfaction or frustration in their teachers.
What do these things have in common? My talk in Europe references Masahiro Mori's theory of "Bukimi No Tani" or the uncanny valley. His theory has to do with the eerie effect of humanoid robots as they approach more human-like qualities. I make the connection to the uncanny valley to animated reenactments affect on jury perception. A kind of unwanted side effect to being more accurate in depicting crime scenes. This eeriness creeps in the robotic interaction with autistic children and can have unwanted responses to their already socially distressed clients. The article described a fix where a staged encounter with the robot interacting with humans (a wave to each other, or short conversation) alleviating fears the child had with the robot. The children quickly became comfortable because of their perceived "humanoid" motion. It is important to note that this is a visual perception of motion and not to human characteristics that can be described as aesthetic. For example, they have human-like hair we find attractive, or human-like eyes either beautiful blues or bodacious browns makes little difference in their comfort level. It is precisely the fact that they react within one and a half seconds of a child's inquiry that makes them feel human.
The other connection to my research into motion suggests we recognize avatar behavior as "artificial" making us more comfortable in their existence possible because we know then avatars are not human. This is contradictory to autistic children being comfortable in aesthetically "odd" or "unusual" robots who physically respond in more human ways. I am not an educator of the mentally disabled but I do know that children who suffer from autism are motivated but hyper-sensitivity to physical response and audio signals. "Treatment may include having the patient handle materials with different textures or listen to different sounds" HealthCommunities.com. The research question arises in what connection to the physical interaction and viewing physical responses performed by robots or even avatars aid therapy in treating autism.