Honestly I don't see the issue at all. So long as your infrastructure can handle it and there is no cost to you it should be fine.

The only proviso is that you should not be able to block distillation on your models you trained using others data.

I also don't appreciate the sensational framing of adult content here, it's purely meant to appeal to irrational emotions. We all know how models train, input and output are very different things here.