Security blogs have been warning of this situation for a long time but don't have the circulation that (eg) symantec has. And a lot of it IS fixable by google.
There are a number of IP blocks which are regularly associated with scam/virus sites, quite apart from the botnets that redirect to these serving sites. It surely isn't beyond google (and others and their detection partners) to check the IP of URLs that are submitted. A simple check of site content on such an IP, using a cloaked bot masquerading as (eg) the most vulnerable browsers (eg IE, Safari), should confirm suspicions.
It would also help if a large company such as google used its clout to pressure the registries to turn off IP blocks of hosting companies that run exploit-friendly servers. This has been done a few times before so it obviously CAN be done again.