Aslan stands triumphant. My prompt: "Lion standing over broken stone table, standing stones and sunrise in background." 

The White Witch on her throne, guarded by wolves. Really like how they picked up on the aesthetic: "Queen in white robes seated on ice throne, wolves on left and right, dark medieval castle interior." A couple good results for this one!

From the Magician's Nephew: 

Aslan singing the world into existence, and Uncle Andrew gloating over one of the green rings. 

Prompts: "Golden Lion singing, night sky with moon, stars and nebulae in background." 

"Greedy old man with fluffy white hair in Victorian clothes holding green ring in his palm, Victorian study bookshelves in background." 

The AI decided Uncle Andrew should have a beard, which is an interesting choice, if not canon with the Baynes illustrations...

"Victorian boy and girl riding on brown winged horse, flying over mountains," yielded a pretty good result, barring Digory's unfortunate lack of a face, as well as the fact that Fledge seems to either have wings on his flanks, or has let Polly borrow them! 

Prince Caspian. 

Prompt: "Boy in medieval armor blowing horn inside cave, light beams coming from above, broken stone table in background." 

The AI decided Caspian only has one arm, and the horn should look like a bullhorn. 🤣

And of course, Reepicheep! (Prompt: "Mouse with gold ring on his ear with red feather on top, holding rapier sword, forest in background." ) The AI objected to the poor mouse being sent into battle without armor, and so remedied the problem on its own...

And Lucy with Aslan turned out near-perfect, except for Lucy's toes. (Prompt: "Girl hugging lion, lying down in pine forest, light beams from above.")

The Voyage of the Dawn Treader! 

For this one, I basically tried to recreate the picture from Aunt Alberta's bedroom. The result was surprisingly accurate - it even added the dragon tail without me telling it to! 

Prompt: "Medieval ship with dragon figurehead and purple sail, greenish ocean waves and cloudy sky in background." 

The Silver Chair! 

Puddleglum according to AI is simply hilarious. 

The gnomes are downright eerie, imo. 

Rilian, on the other hand, is spookily cool. There's no explanation as to why they gave him a hood. Also, the gnomes have considerately provided him with... a chamber pot?

And Jill is very daring while riding Glimfeather. I guess she learned nothing on that cliff..

Sorry  Cleander, but I've seen all I can bear. I just don't like them.

No matter what instructions you give it, AI will look for related pictures and clues, and will turn out some derived fantasy picture. 

By the way, none of Narnia is set in Victorian times. MN is Edwardian. 

Edit: as pointed out below, I am shamefully wrong on the Victorian point.  I concede that MN took place very late in the Late Victorian period.   Almost Edwardian but not quite.

One thing that does interest me about AI derived art, is how much they resemble dreamscapes - not like dreams in the aspirational sense, but in the practical sense of what happens in your brain at night.

As far as I'm aware, dreams aren't all that well understood by psychologists, partly because of how difficult they are to study under scientific conditions, but these AI images appear to follow similar rules in their construction. For example:

• They appear at first to be very life-like and real, and yet once you look at the details it all goes a bit weird. It's like it understands the general shapes and forms of the things it's trying to depict, but doesn't understand their functions, so can end up doing some wacky things.

• They tend to be constructed from vast amalgamations of things you already know from various places - a face here, a location there - and so end up being both eerily familiar, and yet completely alien at the same time, giving you this really weird feeling inside

• The universe the things depicted inhabit appear to have no concept of the laws of physics, or any other governing rules - again it allows for some spectacular visuals, but also some oddly disturbing abominations.

Anyway, all of that is to say that I find the AI images fascinating, even if they are simultaneously amazing, disturbing, beautiful and unsettling to look at.

@coracle I get it! AI certainly has its limits. I was more surprised these turned out even half as well as they did (and amused at some of the odd "creative" choices made). I also did not include the absolutely horrendous results I got for these same prompts - this is the best AI had to offer! Also I totally overlooked the time period issue - fortunately Victorian dress at least at one time would have strongly resembled Edwardian, so hopefully the gap isn't too wide! 

@icarus, You're right about it being like a dreamscape. It looks like something a half- waking mind would do, filling in details rather vaguely and focusing more on getting the general look right. 

Posted by: @coracle

By the way, none of Narnia is set in Victorian times. MN is Edwardian. 

Actually, if we're going to split hairs a bit here... according to Lewis's later timeline (which may not be absolutely canonical), Digory and Polly entered Narnia in our world's year of 1900. Queen Victoria was still alive until January 1901, so the Earthly setting of MN is in fact very late Victorian!!  

That aside, I agree the AI art is interesting and sometimes a bit laughable, but I also can't say I like it. Even artistically and aesthetically, that particular art-bot doesn't seem to have much idea of what makes for a good picture — the distorted faces and bad perspective and so on. I would have thought AI was a bit further along than that by now!!  

But just in general, yes, they're too generic and derivative and contrived and a bit tacky. I'd still rather see artwork from the imagination of a real person, preferably one with some basic art training or at least a really good eye for composition. Although I realise I've been known to complain about human beings' depictions of Narnia too, like the one someone posted recently of Shasta riding alongside Aslan, where Aslan was shown as much shorter than Shasta's horse and with merely a tinge of a golden glow and a vaguely smiling expression on his face, which suggested the artist hadn't actually read the awe-inspiring description of that scene in the book. But at least that person, whoever it was, could draw better than a Microsoft art generator...  

Posted by: @courtenay

Posted by: @coracle

By the way, none of Narnia is set in Victorian times. MN is Edwardian. 

Actually, if we're going to split hairs a bit here... according to Lewis's later timeline (which may not be absolutely canonical), Digory and Polly entered Narnia in our world's year of 1900. Queen Victoria was still alive until January 1901, so the Earthly setting of MN is in fact very late Victorian!!  

Alas, I was wrong! I have added an edit to my post.
[it's good for my humility]

That Puddleglum was pretty good as a character depiction. That Caspian in the cave, however... (shakes head)

The quality of the art depends on whatever generator you use and what it's trained on. The same prompt using different engines creates different stuff. Midjourney consistently gets the best results. Even then, you have to tweak it a lot!

Posted by: @snowangel

Mod note:

The moderators have updated the fan art section rules with new rules for AI art. Going forward one daily post of up to five images is allowed. If more than one post is made in a day or the five image limit is not followed, posts will be deleted.

 

SnowAngel

fan art moderator

Noted, thank you!

Posted by: @the-mad-poet

"Victorian boy and girl riding on brown winged horse, flying over mountains," yielded a pretty good result...

Just out of interest, do we know for sure that Fledge is brown? I don't remember the book specifying that (though it's been a while since I last read MN) — I'm not sure his exact colour is mentioned at all. As his original name was Strawberry, I've always imagined him to be strawberry roan — that is, with chestnut and white hairs intermingled to give a light reddish effect.

Whatever colour he is, anyway, I assume his wings were more sensibly positioned than they are in that art-bot attempt!!