You can do this with your own AI program which is complicated and expensive or Photoshop which is much harder but you can't with openAI.

Making it harder decreases bad behaviour so we should do that.

I don't know the specifics for this reported case, and I'm not interested in learning them, but I know part of the controversy with the grok deep fake thing when it first became a big story was that Grok was starting to add risqué elements to prompted pictures even when the prompt didn't ask for them. But yeah, if users are giving shitty prompts (and I'm sure too many are), they are equally at fault with Grok's devs/designers who did not put in safeguards to prevent those prompts from being actionable before releasing it to the public

But the software knows who she is, and why is software anywhere allowed to generate semi nude images of random women?

Guns don't kill people but a lot of people in one country get killed with guns.