Yes it looks like it's adjusting the port length. (In plain english: some speaker boxes have an intentional hole in them, if you adjust the length of the pathway that sound takes to exit the box through this hole then you adjust how bassy it sounds).

To add a hollow cavity into the plastic part would immensely complicate the design of the moulds (assuming you try and implement the cavity in the same style & orientation of what gluing that bit of wood in achieves). The plastic shells of this speaker look like they've been designed for two-part moulds, which is the cheapest and simplest way of designing a mould. Any internal cavities of the part would require bits of steel mould to be in the cavity during injection, those pieces then have to be removed somehow and that would be a nightmare. Two part moulds can just be clamped & separated over and over again without snagging on anything.

For the walls of a speaker to reflect sound they need to have a density that is very different to the air inside the chamber. As it turns out basically anything fulfills this criteria, even cardboard makes fine speakers (just don't get it wet or poke holes in it). Plastic vs MDF wouldn't matter here acoustically, both are fine.

Bits of particle board can easily be cut and glued by unskilled workers. For business reasons the injection moulding might be getting done at a different place to the final assembly, and the product manager who wants the speakers properly ported might only be in charge of the latter. IDK.

glue applied likely by a machine

I suspect this would be all human assembly. They'll probably have motorised torque-limited screwdrivers and jigs to hold the parts on during assembly, but still human arms doing the work.

In particular: stuffing the white polyester wadding in would be a PITA for an automated assembly machine. Humans are tolerant of variation and bits of wadding blowing away, pre-programmed movement robots are not.