I disagree. Not on the taxing stuff.
There is good reason to use AI and ML in various settings, especially due to climate change.
Where? a good case is remote sensing. We have limited tools and data in temporal and spatial resolution.
And yes, we need more data for various things. I just worked on "how to determine safe water intake bodies for fire fighting planes" and no, this is not trivial to know the depth at every position of a water body worldwide to ad hoc calculate it