This article raises an interesting question:
Every time you are exposed to a virus, your immune system builds resistance to that particular bug. So, why can't we build computers that do the same thing?
Unfortunately, it seems that only academics such as Steven Hofmeyr, Stephanie Forrest and Dipankar Dasgupta are interested in such ideas at the moment but with our computers currently being taken over by all manner of beasties such as spammers, spywares, worms etc the computing community as a whole should start asking hard questions about the ineficiencies in the current security model. I think it is safe to assume that with the exception of possibly unix-based systems, our computer systems are not equipped to survive in todays Internet.
An interesting read is Computer Immune Systems and Architecture for an Artifical Immune System which are research papers jointly written by Stephanie Forrest and Steven Hofmeyr. They are academic papers from which Steven Hofmeyr has developed Primary Response, which, according to eSecurity Planet, is a host-based intrusion prevention system modeled on the human immune system.
The idea of modeling computer systems on human immune systems is interesting to me and I will be keeping an eye on developments.
Update: AMD and Intel have a similar idea which involves allowing processors to stop many attacks before they occur but there is not much detail and I am not sure if it works on its own or in conjunction with Antivirus software - it seems marketed towards Windows XP Service Pack 2.