1. The text prompt used to generate an AI diffusion would not be acceptable for title, description or keywords. Often users include such details as camera lens size, filters, artist names in their text prompts. This might be fine for diffusion services but it's meaningless to Stock Customers who are searching for assets like "little girl in pink dress running through tall grass."
2. We would need to see the rejected asset to give meaningful feedback.
I try to generate the highest resolution possible from the AI service. Depending on which service you use, this can get costly. It may take 100 or more diffusion cycles to get 3-5 decent images to work with. And there's no guarantee that Stock will accept them.
Before submitting, always compare & contrast your work with available Stock inventory. If it's as good or better than what Stock has, submit it. If not, do what I do. Put the duds into a "failed artwork" folder. On rainy days, I print them out, grab my scissors and carve them up into little pieces which I eventually use in other art projects. 😁
Hope that helps.
... View more