"The Forest" by Jason Stewart

Here's a subtitled video of a "Lifelike Expressive Avatars" in a proof-of-concept experiment using computer animation.

Vcom3D and the Laurent Clerc National Deaf Education Center of Gallaudet University are researching and developing a proof-of-concept system for creating and delivering animated stories using the full range of facial expression and body language of American Sign Language, as well as manual signs. Results from this Phase I proof-of-concept will provide the basis for developing an Authoring System and run-time software for creating these animated stories.

I suppose in 10 or 15 years from now we'll have computer animation so life like a human being wouldn't be able to tell from a real person or a computer animated one.