Voice recognition for navigating and ''writing'' notes, ect. I think that would require the use of a lot of raw server power. Android phones uses voice recognition, but I think that it's Google's servers that do the hard work. It would negate many of the uses of a toughscreen.
It could probably work for navigation on the Kindle for simple commands, though.
I just might be completely wrong about this, though. I just saw a programme about a Android phone about a few weeks back.