A portal site which lists search results. Each result contains a list of definitions/entries - each from different instances. Editors from different instances can decide to agree with others at multiple levels (granularity), right down to a definition. It would probably be difficult to agree on paragraphs, but not impossible. Search results with a lot of "agreement" are showed at the top of the result list, while those with with no agreement are shown at the bottom.

The agreement dynamics can already be seen in science between journals, articles, authors, even if it's less structured and formal. There are now search engines that use AI to measure an agreement/disagreement scale between papers (connected by citations).

Obviously, there needs to be some way to validate, track, and mark the "bad faith" instances to push them down or out of the results entirely. And that way has to be based on a combination of expertise and reputation, not on universal vote counts.