Will the first pass on a Turing test come from spammers?
CAPTCHA stands for “Completely Automated Public Turing test to tell Computers and Humans Apart.” Well, OK, Turing test may be a bit strong for what is essentially really hard OCR, but passing a good CAPTCHA shows at least some proficiency with pattern recognition. Now that the GMAIL and HotMail CAPTCHA’s have been cracked*, CAPTCHA’s will presumably become harder still. Eventually, you might get a brief reading comprehension test, or more probably be told to identify a picture. What happens when spammers crack these? Eventually, computers may become indistinguishable from humans in automated tests. Of course that’s a long way from passing a true Turing test, but it wouldn’t surprise me if some advances in AI are made by malware developers.
* Of course, the spammers haven’t really “cracked” these in the sense of being able to do them well every time. They only need to succeed a significant minority of the time for their goals.
I am pretty sure there’s some kind of large-scale project going on in the hacker community. We keep getting mysterious ‘random-letter spam’. Akismet filters out most of it but it is getting better at getting past the barrier.