I personally think too many species are obscured as well. I think birds are a good example, since many aren’t really the subject of poaching/hunting anyways.
On the other hand, I know people advocate for obscuring more herp species than are obscured because some of them are vulnerable to poaching/collecting even if they aren’t officially vulnerable.
I personally think it’s a bit of a tough job for iNat to make these decisions. They originally went with IUCN categories, but it is now a more ad hoc approach where different species can be obscured or not. You can flag specific taxa for curation with regards to their geoprivacy status (see discussion here: https://forum.inaturalist.org/t/why-are-some-threatened-vulnerable-and-endangered-species-locations-being-unobscured/12876/2)
There’s lots of other posts discussing some of these issues including with herps: https://forum.inaturalist.org/t/danger-of-locations-on-inaturalist/6602
How they should apply to invasive/exotic species: https://forum.inaturalist.org/t/how-to-handle-global-conservation-statuses-for-non-native-taxa/7118
Disadvantages in trying to use data for research: https://forum.inaturalist.org/t/access-to-data-for-sensitive-and-obscured-observations/1716/4
etc.
I think the general approach to work from IUCN categories is a good one as it errs on the side of caution (obscuring species that may be in danger). As much as I think iNat’s data is useful scientifically (and agree that obscured and private data are way less useful), it isn’t the primary mission to generate great data. But I can also see a strong case for de-obscuring many species, like birds, as you say here.
It would be interesting to hear from active curators who work with birds about this. One downside to the process of individually flagging species to discuss obscuring/deobscuring is how much work it is, but if you can find a good outside source to follow (like eBird), this definitely wouldn’t be as much work. I am not very experienced with birds, but on the face of it, following eBird’s categories seems reasonable, since there is such a strong community there. eBird seems like they have a pretty strong process for determining which species are sensitive and obscuring those data (source: https://ebird.org/news/sensitive-species-in-ebird). The obscuring seems similar to iNat’s (with a bounding box).