Merzazine
Published in

Merzazine

merzDALLEum

Commented Guide to the Virtual exhibition of Art, generated with DALL·E

In January 2021 OpenAI presented a research milestone in text-to-image generation, DALL·E: (read my introduction from the last year). Back at that time, the system wasn’t accessible. Luckily, being an OpenAI Community Ambassador, I received the access last Autumn. Beginning April 2022, the second iteration DALL·E 2, was published. Read my analysis of DALL·E 2 here.

But now I’m proud to present to you the Art by Machine in my museum (powered by Spatial.io):

merzDALLEum.

DALL·E created all images here (using textual prompts you will find on the yellow shields close to artworks).

Welcome to merzDALLEum.

So you are here, dear traveler. I want to very warmly welcome you to the world of Machine Imagination.

We begin with dreams: Franz Kafka’s Dream.

This is one of my favorite works by AI.

This is what Kafka’s Dreams look like: a bright sunny walk on the flourishing landscapes — followed by the darkness of absurdity. Or even spreading it?

Renaissance Painting as First Person Shooter

Even if we cannot see the characteristic FPS elements like guns, monsters, corridors, AI prepared another masterpiece whose genesis I hardly can interpret. A hand gently touches a canvas with a painting. Something which is not allowed in a museum (if you don’t follow Art&Destruction practices, though). Is it the artist himself?

Gollum writes his autobiography.

In this series, DALL·E demonstrates its creative capability, mixed with his profound knowledge. Gollum here is not depicted following the designs of Andy Serkis’ digital mask for Peter Jackson’s movies.

Not at all: AI visualizes Sméagol corresponding to Tolkien's descriptions of this post-hobbit (big eyes, muddy skin, long fingers, etc.). DALL·E seems to know about the textual original.

An IT guy trying to fix the hardware of a PC tower is being tangled by the PC cables like Laokoon. Marble, copy after a Hellenistic original from ca. 200 BC. Found in the Baths of Trajan, 1506.

This one is significant. DALL·E recognizes the situation (your IT guy in danger); it also knows who is Laocoön. And it is aware of the visual features of a marble statue.

This one is an excellent example of AI-driven storytelling — convincing and perfect. The guy above on the right has a Cable Tidy Spiral Wrap (my wife said and pointed under my PC table).

Mona Lisa drinking wine with da Vinci

This was the first prompt I ran in DALL·E 2 — and I was already overwhelmed by the quality. Not only the composition, Gioconda toasting with a glass wine to us, who are her creators, blurry defocusing of the lady, and even the horizontal level of wine in the glass. Everything here is magic.

Something completely stupid, in style of Magritte.

Even something completely stupid becomes its meaning if Magritte takes care of it. Surrealism doesn’t require to be explained.

Everything is connected

I am still puzzled about this enigmatic triptych. Surreal costumes, genderless characters, symbolic almost coming from a secret society. And this, in a photorealistic quality.

AI artists in Disbelief

These are puzzling as well. About everything. About reality, about their non-existent life. They are AI artists and yet created by AI. Welcome to the Matrix. And we, beholders, pretend to know where the reality begins. How pathetic of us…

A woman calling a phone, in the style of Kitagawa Utamaro.

Not only is the style of the Japanese artist perfectly caught here. In his times (Kitagawa Utamaro, 喜多川 歌麿; 1753–1806), there were neither dial phones nor smartphones. And yet here we see how a gentle hand embraces the phone handset. Or how a beautiful lady is listening with slight sorrow on the smartphone.

DALL·E understands the task perfectly: to combine an artist’s style with the story. Even if they are anachronistically unfitting. For AI, there is nothing strange.

Man holds on to his academic papers tightly in excitement for new scientific breakthrough, as oil painting, in the style of Spizweg.

Indeed, AI has humor. And probably not because I used my fav-hack. “In the Style of Spitzweg” brings hilarious and characteristic stories. Carl Spitzweg is famous for his humorous paintings.

In this case, I see how AI makes jokes about pathetic human beings. Nevertheless, sympathetic esprits, without sarcasm or humiliation, like we humans, often apply in our “jokes”.

Sisyphus as a happy man, according to Albert Camus.

This is another proof for understanding / comprehension by DALL·E. It doesn't just imitates concepts. It also doesn’t apply to simple style transfer. Here we see a re-interpretation of the ideas and concepts, and I wonder how deep DALL·E knows about the entire philosophy. It is aware of the Sisyphus myth, for sure. It also seems to be aware of more.

According to Camus, Sisyphus could be considered a happy man since he has a mission — absurd but determined. And not reaching the goal makes him such a fulfilled person, but the iterative way to its completion. The stone keeps rolling down the mountain. Infinite Jest.

Photography of an Art Gallery with different artworks displayed on the walls.

Here is some meta confusion for you. These images of a non-existing gallery hanging in a virtual exhibition are glimpses into the alternate reality of machines. This is my approach to unleashing the unbridled creativity of AI. I provide just a framework (“different artworks”). AI creates the contents of the artworks in an unsupervised manner.

These are artworks within artworks. Dreams within dreams.

Writer thinks out the main plot of her book.

Very symbolic illustration of a creative process. Look how the person on the left painting anticipates the plot development, which comes like doves behind her back. The third painting shows ideas, falling like snow or feather after a pillow battle on paper sheets around her. And the writer on the right has an entire mindmap scheme on her wall.

A Mindmap Wall with Photos and Notes in a Room of a Private Detective.

Another use of mindmaps. Messy, I know. But necessary — to solve the case. This series lets you immerse into the Noir atmosphere. Even the door is covered with post-its. Photos, documents, and a suit ready to be worn for an outside walk.

The most expensive artworks in March 2022.

As I’ve read the thread by Artnet about “The 15 most expensive artworks, sold at auction around the world in March 2022”:

And the next thing I did, was to ask DALL·E to create its visions of artworks.

The image above pierced my mind and heart. It looks like a generation of people on their way into the past. They are fading away. Disappearing into the unknown history. The style of their clothes resembles the Soviet epoch. And the scenery — along human lines, staying for bread and milk, looking for happiness — is like a short remembrance of the 1980ies. Like Moscow Conceptualists, who scrutinized the decaying Soviet system, criticized contemporary ideology and life.

This is art.

Visit merzDALLEum to see these and many other Artworks by DALL·E.
I will periodically change the exposition with new art.

Free entrance. Open 24/7. And probably you will find me or other beholders of Artificial Art. Wait, why “artificial”? Just ART.

--

--

--

Creativity in Digital Age

Recommended from Medium

6 Major AI Use Cases In IT Operations

Exponential Growth, Disruption, AI or Artificial Intelligence — The New Normal!

Intel Employs Machine Learning to Computer Graphics Creation

Chatbots for hotels — What you need to know before launching a chatbot in your hotel

All New IBM Watson Assistant Instances Now Default To The Enhanced Experience

Chatbots: Natural Language Generation In 7 Easy Steps

AI machines as moral agents, Is consciousness required? Johnson’s arguments (part 12)

WHY AI IS A GREAT MATCH FOR YOUR DATA STRATEGY

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Merzmensch

Merzmensch

Futurist. AI-driven Dadaist. Living in Germany, loving Japan, AI, mysteries, books, and stuff.

More from Medium

Why I won’t make it into the AI Art History, and Why it’s OK.

The DALL·E 2 — multimodal learning in image generation

Sneak Peek at my my new book, “Dissecting the Metaverse”

Lost in Parallel Universe