I'm aware of the training problem, openAI got sued by Shutterstock for it. however, we as a moderately obscure meme generator with social aspects have little say in the way the companies manage their scraping.
best we can do is just ask people to allow others to find the AI.
(also, technically the AI's art is "original" the algorithms used simply look at its training data to figure out what kind of thing you mean, then making its own creation based on that.)