I honestly never anticipated that kind of concern. Huh. Well, now I can! Thank you,
@david_wendelken
The concerns are very real.
Ent just finished co-authoring an article for a magazine, and for the article Ent had an AI image generated showing
Tolkien sitting at a desk, in a study room, depicting him "considering every word" as he wrote - a somewhat key element of the article.
Before we could submit the article with this picture, research needed to be done to see if we could use it, given its portrayal of a 'very obvious simile of
J.R.R. Tolkien.'
Ultimately, research showed we could because: 1) - the image was a "likeness" constructed from no existing picture, a real study room, and contained no
books, typewriter, etc. known to be duplicates of 'real' Tolkien things.
However, for safety, it was recommended we:
1. Put a caption on the picture indicating it was AI done.
2. Keep record of the content of Ent's request to AI forming the creation of the picture and its content, engaging AI to construct the picture.
3. Keep record of any requests for change to the picture following the initial creation.
4. Make sure the records were dated, so we could always show exactly when the engagement of the picture was done.
5. Be able to show the picture's integral connection to the article it was created for. (Very simple, in this case.)
THEN, we needed to understand while we had the 'rights' to the picture without infringing on the AI producer itself, AND while we could use the picture, we did not have "exclusive rights" over the picture's use by others in the future, just by reason of having had it generated consistent with Ent's instruction.
Publication is a tangled web.
Here is the work finally submitted with the article. (Entirely based on the INITIAL background request/details Ent provided for its rendering... no changes or alterations needed..! Ent was impressed..!)