It hasn’t been that way for a very long time. In pre-API 2.0 IDEs, yes, that is true.
The namespaced framework (the one where everything was prefixed with Xojo.) introduced native methods for creating and parsing JSON, and that’s where the Dictionary / ParseJSON system comes from. Shortly into API 2.0, the JSONItem class was updated to be a wrapper around that system because of how obtuse and difficult to use the Dictionary system is.
They now do the same thing, while JSONItem handles a lot of the boilerplate and difficult aspects for you, making for cleaner, more obvious, and easier to understand code.
Since Xojo’s target audience is mainly users who prefer easy, I think recommending simple self-documenting code is a better direction.