There is certainly some suggestive evidence out there, but whether that justifies the suspicions about “most people” looks quite dubious. There is a forum post in a thread about “speed identifying” that I found quite disturbing. I find it hard to see how aiming to do fifty thousand IDs in a month can be good for either the site or the individuals involved. Identification isn’t (or shouldn’t be) purely a numbers game. Although there may only be a small number of people doing this, I think iNaturalist has an obligation to explicitly discourage it (even if it’s well-intentioned). There is no problem with gamification per se - but, as with so many things in life, it needs to be done in moderation.
So I personally think the changes are aiming in the right direction, but the specific steps taken so far are probably not the best ones. However, maybe the most effective way to prove this is to treat it as an experiment and thereby show that it either doesn’t solve the problems it is supposed to, or has too many negative side-effects to justify continuing with it.