Put a finger on the scales, show us how it’s done.
Modeling this behavior, without tilting the scales completely is perhaps impossible.
I’m interested in seeing how such code could work, while scaling and not creating artifacts. I can’t think of anyway for such code to be deployed without breaking search and letting attackers gain even more
I totally agree, it could be as simple as the algorithm finding more usages of 'gun' and 'rights' on the Gun Control page than on the Right to Bear Arms page and basing its decision solely off that. However it could be as complex as skewed data in a deep learning model that takes in every users search term and subsequent selected result. No one really knows exactly how its decided except Google's own engineers.
The fact that most other, very similar search terms provide the expected results suggests to me that it is not an intentional misdirection, otherwise the other search terms about guns would also show the Gun Control page.
https://en.wikipedia.org/wiki/Gun_rights is a valid link, and seems like the obvious wikipedia page to deem relevant for that phrase. And Bing gets it right.
I have no reason to think it's some dark conspiracy, but it's certainly a pretty silly failure of their vaunted algorithms. Are there a lot of searches for common phrases where a Wikipedia page exists and Google returns a different Wikipedia page as the top item?
It's a bit like searching for "USA" and getting a result for "Canada"; just because the canonical wikipedia URL is "United_states" doesn't mean that a page about a different country is more relevant. :)