That policy still seems clear as mud if you're using, say, generative AI within Photoshop to help design components of a cover design. I'm also not sure how that's really different from using CC or public domain imagery as part of a design.

"AI-generated: We define AI-generated content as text, images, or translations created by an AI-based tool. If you used an AI-based tool to create the actual content (whether text, images, or translations), it is considered "AI-generated," even if you applied substantial edits afterwards."

My reading of that is that, if you use AI to create any image even in part, that's AI generated.

WTF is the difference between AI-generated and AI-assisted. Seems like a very murky distinction.

I'll give you an example from my recent work.

We just released a new app onto the Apple App Store.

In order to do this, we need to submit a bunch of screenshots, preview videos, and text.

The images and videos, we hand-crafted. A lot of folks automate this, but I think that it's better to hand-craft, if possible, because we need to ensure that specific contexts and whatnot are reflected, and that user privacy is respected.

For the text, we developed a fairly dry blurb, and then I asked ChatGPT4 to rewrite.

I hit "refresh" a few times, and wrote up a text document with the results, and shared that with the team.

We then picked the parts we wanted, and wrote some of our own.

None of the ChatGPT-generated summaries was sufficient, on its own merits. We had to cut-and-paste (sometimes changing the order, and changing to make sure that there was consistency), and add our own content, as well.