I need help understanding how to think about an AI ethics question.
I don’t have answers.
If you do, please leave them in the comments.
I recently saw a demonstration of an impressive AI image generator. As a small part of the demo, it replaced a women’s blouse with a string bikini.
It started with a G-rated picture of an adult and created a G-rated picture of an adult.
When is this OK?
When is it not OK?
Why?
Here is the scenario I am thinking about. It centers around a green sweater.
I properly obtain a G-rated picture of an adult human.
Perhaps it is a movie star, politician, or public figure.
Perhaps it is a friend, family member, enemy, or stranger.
Perhaps it is a fictional character.
I create a version of them wearing a green sweater.
That person would never be caught dead wearing green. Or a sweater. And definitely not a green sweater.
Perhaps I know this. Perhaps I don’t.
Perhaps I am doing this because I think they would look wonderful in a green sweater.
Or perhaps because they would hate it.
Or maybe I just have a thing for green sweaters.
Perhaps I make it with AI.
Or pre-AI Photoshop.
Or a paintbrush on a printed copy.
Let’s presume for the moment that I am not going to share it with anyone, as sharing it raises issues of its own. We an even presume that I never mention its existence and it is never stolen.
Is it OK for me to create that image?
Which of those “perhaps’s” matter?
And given a situation where it is not OK for me to create that image, is it OK for me to just imagine it?
Part 2 — The Really Scary Part!
The below image was my initial thought for cover art.
It’s G-rated.
It’s not a real person.
It’s not even a person.
Still, it was well outside my comfort zone to use it as cover art.
And barely within my comfort zone to include it all.
Would you have been uncomfortable using it as cover art?
Why or why not?
Something else you might use the comments to enlighten me on.
