In a recent example, Stable Diffusion, an open source AI system that generates images from text prompts, was released with a license prohibiting certain types of content. But it quickly found an audience within communities that use such AI tools to create pornographic deepfakes of celebrities.
This is a big question, whether use restrictions, which are becoming prolific (RAIL license, for example), can be enforced. If not, and that's a big if, it might create a situation of "responsibility washing" - licensors can argue they did all that's possible to curb harmful uses, and these will continue to happen in a gray / dark zone
