Just read on The Register that spammers seem to have adopted sophisticated software – possibly an OCR tool – to break through protections laid by web e-mail providers such as Yahoo!, Microsoft and Google to avoid automated account creation.
I think this begs the question: will the arms race between malware writers and security companies drive future developments in AI? Already there are new developments (such as Process Query Systems) aimed, at least partially, at improving network security; and the spammers certainly seem to be doing their own applied research. Could this be the beginning of the pragmatic, sustained AI industry that the industrial-military complex failed to bootstrap?
