At the same time. This is a clear "why not both?" situation.

Let's care for bees. Of course. But engineering even for it's own sake is beneficial.

Some AI problems (or really NN problems) are stupidly difficult. Recognizing individual flower parts from a remotely driven camera on a small copter for one has applicability to about every journal even adjacent to aerospace, control systems, and probably distributed control and consensus. That shit drives science too. Physics informed loss function reduction (for PINNs) are super cutting edge and is at the intersection of science and engineering.

My aero research lab that worked on military systems and airports precipitated a cool as hell line of research into the spread of feline diseases using overlapping principles.

It's all good stuff. As long as those copters don't run on ground up bees, I think it's cool someone is getting 6 or 7 figures for a group to research it.