I don’t think Asobo has inhouse AI team for the tasks people here suggested to do. However, MS does have strong AI departments that could definitely support the endeavor if they are willing to invest more.
As for volumetric water, I think it could be possible when there’s additional computational power to be leveraged, or there’s a demand for water-based simulation (more seaplanes, boats, etc.)
My wish is that they could open up APIs for
- auto-gen (modders could implement different AI algorithm with different texture lib customized to each region)
- post-processing for satellite imagery (color correction, algorithm based water masking, etc.)