Generative image AI as reading companion

A friend recently mentioned that he is feeding DALL•E prose from the fantasy novels he is reading, and asking it to create images of the scenes. This sounded like an excellent idea to me and since I am currently re-reading Lord of the Rings (this beautiful edition), I’ve been doing the same. It has genuinely enhanced my enjoyment of the book to create “text-accurate” images of the characters and scenes as I read.

Here are a few of my favorites (yes, I’m still in Fellowship)…

Tom Bombadil exactly as he appears the first time in the book:

Aragorn the first time he is fully described:

Frodo when he encounters the Ringwraiths on Weathertop. I forgot about the red sword (since the movies gave their swords a blue glow):

In case anyone is interested, here is the prompt I’m using:

Please generate images of the description below. Create one image in each of the following styles:

  1. Surrealistic Photorealism: Combining hyper-realistic details with surreal elements to create a dreamlike yet sharply detailed image.
  2. Gothic or Dark Romanticism: Emphasizing darker, more mysterious elements of the description, with a focus on emotion, nature, and the macabre.
  3. Magical Realism: Blending realistic and fantastical elements in a way that treats the extraordinary as everyday, enhancing the mystical aspects of the description.

I then choose the style that seems most appropriate, and sometimes add/remove some details. It’s a fun experiment, you should try it!