The question of what should be obscured and what happens when something is obscured are two separate ones.
Right now the answer to the first is if it has a conservation status, it is obscured. In some parts of the world the community can agree to turn on or off the obscuring in other places changes arw forbidden. Which leads to the ludicrous situation we have in my home province of the single most observed species in the province being obscured based on threats 5000 kilometers away, yet far less common members of the family which share the same threats in the province here are open.
My primary concern is that no single user should be able to change the obscured status of a species and most importantly no user should have access to any functionality that overrides the obscuring or an observer’s decision to obscure something.
Unless or until inaturalist designs some program to validate credentials and how to staff it etc then the current system should stay.
None of your examples are 100 percent accurate.
A user wanting to do a project would not need to reach out to the owners of all 974000 observations in Canada because a high percentage of them are open not obscured.
Joining a project is not the only way to share obscured locations, it can be done via the trusted users functionality.
If a researcher in Canada wants to do work on elephants they dont need to create their own project. If there already is one, they can ask the admin to be made a co-admin which grants them access or if they will share the data they have access to. Presumably the cohort of elephant researchers is small enough to recognize legitimate members.
If someone right now feels they need the exact location of each of those 594 elephants and there is agreement that elephants should be obscured, then no user should be able to self appoint themselves into a position to overrule that. Period. Full stop. They can decide if the work involved to follow the process to get it is worth it.