A source at Wizards of the Coast has said that Shkipin’s use of AI art was not something that the team was aware of. Further, they state that no text in the book was AI generated. Wizards has said that they will update their guidelines to more explicitly prevent these sorts of incidents from happening in the future.
io9 has reached out to Ilya Shkipin for clarification and comment. Shkipin has further stated on social media today in light of his confirmation and the deletion of his prior posts that “the future of today[s] illustrations is being discussed.”
I don’t see anything in here about them removing the art.
I believe Shkipin confirmed in a tweet that the illustrations would be reworked. No idea if this was caught in time to change the first printing, or if it will only be changed for the second run forward. Also no word as to whether or not they will be reworked with/without AI. I suspect they will, given how prolific Shkipin is in the AI/NFT front, but I hope WotC will take a closer look at things to prevent that from happening.
I had taken the artist’s statement that the art is being reworked as evidence of the art being pulled, and at least future published books not containing them.
https://twitter.com/i_shkipin/status/1687829743268442112
I should have waited until the actual statement (linked below), as it’s clear that the artist was making something of a lie of omission above, as I believe he purposefully wanted readers to believe that the art was changing in the books, not that he would redo the art to… essentially no promised effect. Maybe he was lying to himself with the hope that he could get another go at?
https://twitter.com/DnDBeyond/status/1687969469170094083
I’ve updated the title of this post to reflect the info from the official WotC statement on the issue. I’m glad they won’t be allowing AI “augmented” submissions from their artists going forward, but I would only be truly happy with a statement that they will be taking the massive inconvenience of making sure that the art is revised for future publications at a minimum (I think a recall would be asking a lot tbh).
