Licensing is not a simple semantic problem. It is a legal problem that have strong ramifications, especially things are on their way to standardize. What Facebook is trying to do with their "open source" models is to exhaust possibility of fully open source models to be industry standarts. and create an alternative monopoly to Microsoft/OpenAI. Think of it as if an entity had right to ISO standards, they would be extremely rich. Eventually researchers will release pretty advance ML models that are fully open source(from dataset to training code) and Facebook is trying to block them even before start to prevent of the possibility of this models to be standard. This is a complementary tactic of the industry to closed source rivals and should not be understood as challenging to them.

A good wording for this is "open-washing" as described in this paper: https://dl.acm.org/doi/fullHtml/10.1145/3630106.3659005