in this thread, i don’t want to go off on too much of a tangent discussing the merits of copyright on iNat observations. there are other threads that go over this, and here are my thoughts on one of those threads: https://forum.inaturalist.org/t/does-copyright-law-allow-you-to-restrict-use-of-your-observations-the-data-this-is-not-the-photos/41709/5.
each repository / organization can decide whether and how it wants to honor user licenses on observations, and each repository / organization can protect – or not – the data stored there using various mechanisms.
in the specific case of iNat and GBIF, it seems like GBIF won’t accept data unless it’s licensed CC0, CC BY, or CC BY-NC, iNat has decided it won’t push observations to GBIF unless licensed CC0, CC BY, or CC BY-NC; and iNat has chosen to ask its users to pick the license for their own observations rather than applying a blanket license.
iNat won’t actively push the data to such a repository, but if such a repository decides to scrape data from iNat, iNat isn’t going to stop it simply because it’s going to disregard observation copyright / licenses.