Hello Gary,

I'll start by answering your last question. Yes, I would love use interactive avatars with children; especially those with autism. It would be fun in many ways for all involved.

Right now there is strong circumstantial evidence to justify use of interactive avatars to teach children. Probably the strongest support are results from other forms of video-modeling. This includes, adult/peer modeling, point-of-view modeling, and self-modeling. Research on interactive avatars, animation, and virtual reality is still in its infancy. Successful outcomes are almost universal with these methods and that includes long-term maintenance and generalization to other skills and behaviors. 

Another area that indicates that interactive avatars will be successful is the medium itself. As McLuhan said "The medium is the message" and he has been proven correct. One indication of the power of video is the recent inclusion of "gaming addiction" into the Diagnostic and Statistical Manual of psychological disorders (DSM-5). The power of video is so strong we must be careful not to get carried away with it. But used judicially, this medium can be a powerful teaching tool.

We get further support, specifically for children with autism, from the work of Temple Grandin and researchers looking at learning styles. In her book "Thinking in Pictures" Grandin relates how she learned as a child and student. She spontaneously translated incoming language information into images ("...like a VCR running continuously in my head"). Although not all children with autism learn like Grandin, there is evidence to suggest that they have a tendency to be visual learners (e.g. Bauman, 1999). Grandin suggested another reason why video would be effective for children with autism. These children have difficulty with social interactions with others, by definition. This can be one of the greatest challenges in the student-learner relationship. The use of video seems to negate this problem. Grandin attributes this to the lack of social obligations required by screen engagement. It appears nonthreatening to them.

This brings us to interactive avatars. The other forms of modeling are largely passive. It is an observe and learn experience. We don't need a lot of research to conclude that animation is very attractive to children. I imagine Disney and Pixar have done marketing surveys that would support this. Invirtua's interactive avatars (which are "pixar quality") provide the opportunity for children to talk directly to the animations. This could be a very powerful feature. And, then there is the next step where the child actually animates the images. We know that teaching can be a very effective way to learn and that's what will take place when the child has control. Finally, there is the emotion manipulation that covers every possible facial  configuration. There is research that indicates that animations can be used to teach emotion recognition.  Once again moving this instruction from passive observation to child-control could be very powerful and empowering as well.

Excuse the double negative, but I cannot see this system not working. The circumstantial evidence is too strong and too closely tied to aspects of the interactive avatars.