Artificial Intelligence Images

These days (February, 2023) there has been a great deal of thoughtful discussion and worried chatter about the nature and possible effects of artificially-generated texts, but not so much about computer-produced pictures and music. So I thought I might post some of my (actually, DALL.E-generated and Photoshop enhanced) images as a starting point for developing my own emerging ideas about AI innovations in our time.

An early effort, using the stimulus words BAR MUSIC NIGHT:

(Click the image for a LIGHTBOX view)

I confess that I have no understanding of how DALL.E generates these pictures. I used the same stimulus words several times, and in each instances the resulting pictures were quite dissimilar. What I especially appreciated about the image reproduced above is both the ambience of the scene and the sensation of looking over the shoulder of someone standing in front of me.

Next I thought I would try paring an actual photo taken by me in a a Flagstaff, Arizona cafe at night. For the AI image on the right I put in the stimulus words BEATNIK POETRY READING COFFEE SHOP:

(Click the image for a LIGHTBOX view)

Once again, the AI-generated image on the right is the one I chose from many less successful iterations. It had some flaws — a crude scar on the man’s face and an imperfect blotch below the woman’s nose — both easily corrected in Photoshop. But what especially liked was the contrast between the intimacy of the photo and the “together but separate” picture of the couple on the right.

“WISHING — OBSTRUCTIONS —HOPE”

(Click the image for a LIGHTBOX view)

Then I tried a three-stage development story depicting spiritual advancement. The images of each stage were chosen among the many unsuccessful attempts by DALL.E to give me what I wanted. All of the images, however, were quite different in appearance (color, perspective, angle of view, etc. I added a Photoshop silhouette of the woman on the first panel to give the “story” a human context.

Night Thoughts at Dawn

Dawn after Desolation

A prediction about the narrative capabilities of the computer:

“The computer looks more each day like the camera of the 1890s: a truly revolutionary invention on the verge of putting to use as a spellbinding storyteller." 

— Janet Murray, Hamlet on the Holodeck (1997)

Still Pictures to Moving Pictures, Computer Information to AI Creations

 One of the — perhaps THE — most momentous in the history of the visual arts was the metamorphosis of photography into motion pictures in the late nineteenth century. It took half a century, from the invention of stillmphotography in 1839 until the first moving pictures in the late 1880s for some inventors discover that if they ran still photos on celluloid in front of a light projector and sent them to a screen, the result would convey the illusion of motion.

Then came sound, the jump-cut, the ability to record dramas of all kinds and allow them to be repeated on demand wherever there was electricity to power the projector.

Some 50+ years after the advent of vacuum-tube computers, entrepreneurs have started generating text, images, and music with comparable results. Some folks are using movies as a basis for producing … well, I don’t have a name for what results. It could be as far removed from traditional cinema  as early black-and-white photos are from Jackson’s Lord of the Rings trilogy.

The first venture into the realm of computer-composed compositions came in the field of music/ Here is David Cope’s version of a Mozart “Rondo,” generated by his computer, “Eliza”:

https://www.youtube.com/watch?v=hdHLGqDWpPo

For a brief glimpse of what might be the future of AI-generated “movies,” check out this brief essay:

https://thenextweb.com/news/text-to-image-generator-creates-award-winning-ai-film-the-crow

Be sure to watch the “Painted” and “Crow” videos.