two problems

1. Is it possible for a remote server to prove that the software it runs is what it claims to run to the client.

2. Is it possible for a piece of software to operate on encrypted content in such a way that the owner of the machine cannot know what the software is doing, and no information about the data being processed is exposed?

>>14 encrypted instruction sets come to mind, but i haven't really looked at how they work.. Probably just another feature for viruses..

Maybe if the server's memory was completely open and observable (read only) through the network you'd have half a chance..? if All operations could be verified..