Short answer: No.
Longer answer: No way.
From a theoretical point of view, it's a Turing-decidability problem. It's imposible, in principle, if an email is "unwanted" just by looking at it. You have to understand it, and that takes us a level deeper into machine intelligence that we would be able to code.
On the other hand, a inmune-system like approach would probably work, at least from the receiving side. As soon as somebody receives an unsolicited email, some kind of antibodies
. That was proposed some time ago in an Usenet Posting
, and, recently, paper by canadian researchers (PDF)
delves deeper into it.
However, it's not as simple as it looks. Spam messages mutate, so it's difficult to create a unique signature that can be transmitted as an antibody
(this is theoretically called a quasispecies
); it does not come from a single source, so it cannot identified by vector.
Are other solutions possible? I'll give short and long answers in the next postings.