http://www.theregister.co.uk/2011/03/08/ipv6_spam_filtering_headache/
Perhaps people will move to whitelists instead of blacklists.
It's also going to present problems to forums in terms of preventing spammers.
All it means is that instead of blocking /24s or similar for IPv4, they'll be blocking /64s or similar for IPv6.
Allocations will either be /128 to single end user PCs, or /64 or bigger to individual customers. (Larger customers will probably get a /48.) Blacklists will continue to work just as they have to date. That none of the DNSBLs are currently designed to handle IPv6 is more of a problem. When there is proven need, undoubtedly they will.
I run an IPv6-reachable mail server (/128 allocation from Gandi, not a tunnel). It does SMTP with optional TLS, imaps, and the only way that email can be sent through it is using a login over SSL on the submission port. I use several IPv4 DNSBLs to block connections and a final classification of spam/ham using spamassassin. My widely known primary email address gets maybe one unfiltered spam a month, but I can't as easily filter my @debian.org mail which gets forwarded on from Debian's servers, so I get about three a week by that route.
I don't use greylisting (but that would work equally well for IPv6) as I found that it is a resource hog that doesn't get you that much more protection.
Up to two years ago, I ran an ISP for 14 years. I now work in a different field of computing.
Else some other safe system will be applied. Cachpa for mail? Would stop bots at least to some extent.