Websites with a ForgiveAndForget policy, like wikis with KeptPages, are vulnerable to having their copyright swizzled out from underneath them in payment for their good natures. Consider a wiki like WikiPedia whose articles are very general, generic prose written to avoid idiosyncracies specific to Wikipedia. It's possible for someone to take a copy of a page from there today, wait some months for the KeptPages history to expire, claim the original copyright, and then sue Wikipedia for "stealing" the material.
One response is to eliminate the ForgiveAndForget policy (as Wikipedia did long ago). That is, to keep every version, every flamewar, every mistake, and every liability. That is, if someone libels today, and you version everything, that libel will be recorded forever. Then you get into deleting individual versions, and that's another major problem. Who has the power to do this? Do you trust the administrators? Here, the administrators don't trust themselves, which is why PageDeletion is not an administrator action but a communal one.
Right now maybe the best answer is to pray that the WebArchive can corroborate your copyright, even though the WebArchive essentially erodes your right to ForgiveAndForget. Maybe there is no answer. But whatever you do to solve this problem yourself, avoid the old KeptVersions idea. The key to scalability is ForgiveAndForget.
CopyrightRegistration is the obvious solution to CopyrightSwizzling. It costs $30 every 3 months.