It seems to me that AI is suggesting this specific ID for a lot of pupae which is why so many end up as D. procinctus.
Yes, this is exactly what is happening. As someone who IDs Lepidoptera, I have also noticed widespread misidentifications of the generic looking, brown, moth pupae. At some point or another, by chance, Dargida procinctus must have been suggested by the CV, unwitting users selected it without thinking twice, and now we have a big misidentification problem that feeds into itself. Dargida procinctus, along with a handful of other random moth species, have now become garbage bins for all these generic, brown pupae.
Should the pupae just be left with the suggested ID or moved to, say, Owlet Moths and Allies?
Please, if you have the time, try to move them out. The only way to combat the CV is to keep bumping the misidentifications up to a higher level. In most cases, there really is no way to identify moth pupae past the family level except for some large, and morphologically distinct species like Sphingidae and Saturniidae. A lot of times I just keep it simple and go all the way back to Lepidoptera.