Who is the artist?

December 02, 2022

I have been brainstorming on my own for a few weeks now on how to use text and image models to make an art piece. I had some ideas like using Sol LeWitt’s instructions but got stuck halfway. Today, feeling really uninspired I tried to brainstorm with ChatGPT a bit directly.

To my great surprise, ChatGPT had great ideas and here are the results. In the end, I asked ChatGPT, if it thinks itself as artist or me. ChatGPT replied:

As an AI language model, I am not capable of creating art myself. Instead, I provide suggestions and ideas for art projects, but it is up to the user to decide how to implement those ideas and create the final artwork. In this sense, the user is the artist, and I am simply providing suggestions and inspiration for their work.

I would beg to differ.

This artwork is entirely based on a conversation with ChatGPT, and feeding the idea into Midjourney’s text to image generation system. I did the overlay of text manually based on ChatGPT’s suggestions.

image
image
image
image
image
image
image
image
image
image
image
image

The full transcript of the exchange below.

Maximal Margin:
What's a good art idea with generative text and image models?

ChatGPT:
One idea could be creating a generative art piece that uses a combination of text and image models to create unique, abstract landscapes. The text model could be used to generate random phrases or words that are then incorporated into the image, such as using the words as the names of various geographical features in the landscape. The image model could be used to generate the actual landscape, incorporating elements such as mountains, forests, and bodies of water. The end result would be a unique, one-of-a-kind piece of art that blends text and image together in a visually striking way.


Maximal Margin:
Okay. That's interesting. Can you connect to some abstract artist's styles?

ChatGPT:
One artist who often incorporates text and abstract imagery in their work is Jenny Holzer. Holzer's work often features bold, impactful phrases and statements written in large, bold fonts, which are then incorporated into abstract landscapes or compositions. This combination of text and abstract imagery is a common feature of Holzer's work, and could be a source of inspiration for the generative art piece described above. Other abstract artists who incorporate text in their work include Barbara Kruger and Lawrence Weiner, whose work often features provocative phrases and statements presented in a bold, graphic style. These artists could also serve as a source of inspiration for a generative art piece incorporating text and abstract imagery.


Maximal Margin:
okay. that's nice. how do we get started?

ChatGPT:
To get started on a generative art piece incorporating text and image models, the first step would be to gather a collection of images and phrases to use as the input for the models. This could involve taking your own photographs or finding images online, as well as brainstorming a list of potential phrases or words to use in the piece.

Once you have a collection of images and phrases, the next step would be to select and train the generative models that will be used to create the art piece. This could involve using a machine learning platform such as TensorFlow or Keras to train a model on the images and phrases, or using a pre-trained model and fine-tuning it to work with your specific inputs.

Once the models are trained, the final step would be to use them to generate the actual art piece. This could involve using the text model to generate random phrases or words, and then incorporating those into the image model to create unique, abstract landscapes. The end result would be a one-of-a-kind art piece that blends text and image together in a visually striking way.


Maximal Margin:
sg. can you give me some random phrases or words to start with? i want a modern, stern, grey concrete, berlin vibe but cleaner and lighter.

ChatGPT:
"Neon skyline"
"Concrete jungle"
"Urban decay"
"Industrial landscape"
"Sleek and modern"
"Cold and austere"
"Brutalist architecture"
"Asphalt and steel"
"Futuristic metropolis"
"Gleaming towers"

Maximal Margin:
maybe more details and concept? and longer?

ChatGPT:
One idea for a generative art piece incorporating text and image models could be a series of abstract landscapes that explore the contrast between the cold, industrial aesthetic of a modern city and the warmth and vibrancy of its inhabitants. The text model could generate phrases that reflect this contrast, such as "Neon skyline" and "Concrete jungle," while the image model could generate abstract landscapes featuring towering buildings, gleaming towers, and sprawling cityscapes.

The overall concept of the piece could be to capture the tension and contrast between the industrial and the human, using bold, graphic phrases and abstract imagery to create a visually striking and thought-provoking series of artworks. The color palette could be a mix of cool, muted tones for the industrial elements, and warm, vibrant hues for the human elements, creating a dynamic and visually interesting interplay between the two. The end result would be a series of unique, one-of-a-kind artworks that explore the complex relationship between the built environment and the people who inhabit it.


Maximal Margin:
okay working on it. other ideas? i like edward hopper and want to draw inspiration from him

ChatGPT:
Another idea for a generative art piece incorporating text and image models could be inspired by the work of Edward Hopper, who is known for his paintings of urban and rural scenes that capture the mood and atmosphere of everyday life. For this piece, the text model could generate phrases that reflect the solitude and isolation often present in Hopper's work, such as "Lonely streets" and "Empty rooms." The image model could then generate abstract landscapes or cityscapes that incorporate elements of Hopper's signature style, such as the use of light and shadow to create a sense of atmosphere.

The overall concept of the piece could be to explore the theme of solitude and isolation in the modern world, using a combination of text and image to create a series of artworks that capture the mood and atmosphere of everyday life. The color palette could be muted and subdued, reflecting the mood of the piece, and the composition could be focused on the relationship between the built environment and the people who inhabit it. The end result would be a series of unique, one-of-a-kind artworks that explore the themes of solitude and isolation in the modern world.


Maximal Margin:
nice. got the images generated. what fonts should i use and color for these text overlays

ChatGPT:
For a generative art piece incorporating text and image models, it would be a good idea to choose fonts that are bold and impactful, to help the text stand out and draw the viewer's attention. For the phrases generated by the text model, a sans-serif font such as Helvetica or Arial could be a good choice, as these fonts are clean and modern and would complement the abstract imagery well.

For the color of the text overlays, it would depend on the overall aesthetic of the piece and the mood you are trying to create. For a piece that explores the contrast between the industrial and the human, as described above, a cool, muted color such as grey or blue could be a good choice for the industrial elements, while a warm, vibrant color such as red or yellow could be used for the human elements. This would create a dynamic and visually interesting interplay between the two, and help to highlight the contrast between the two elements.

Overall, the key is to experiment with different fonts and colors to see what works best with the images and phrases generated by the models, and to create a cohesive and visually striking overall aesthetic for the piece.


Maximal Margin:
i also want to pay homage to louise bourgeois. can i do that?

ChatGPT:
Yes, it is possible to incorporate elements of Louise Bourgeois' work into a generative art piece incorporating text and image models. Bourgeois is known for her sculptures and installations that explore themes of sexuality, identity, and the body, often using abstract forms and organic shapes to create a sense of unease and discomfort.

To incorporate elements of Bourgeois' work into a generative art piece, the text model could generate phrases that reflect her themes and concerns, such as "The body" and "Identity crisis." The image model could then generate abstract landscapes or compositions that incorporate organic forms and shapes, as well as textures and patterns that suggest the human body.

The overall concept of the piece could be to explore the theme of the body and identity, using a combination of text and image to create a series of artworks that capture the mood and atmosphere of Bourgeois' work. The color palette could be muted and subdued, with earthy tones and organic hues, and the composition could be focused on the relationship between the abstract forms and the human body. The end result would be a series of unique, one-of-a-kind artworks that pay homage to the work of Louise Bourgeois.


Maximal Margin:
cool. any last details i should add on?

ChatGPT:
One final detail to consider for a generative art piece incorporating text and image models is the placement of the text within the images. Depending on the aesthetic and concept of the piece, the text could be positioned in a variety of ways to create different effects. For example, the text could be overlaid on top of the image, with the words appearing in front of the landscape or cityscape. Alternatively, the text could be integrated into the image, with the words forming part of the landscape itself, such as the names of geographical features or buildings.

Experimenting with different placements and arrangements of the text within the images can help to create a unique and visually striking aesthetic for the piece, and can help to enhance the overall concept and meaning of the artwork. So, it is worth considering this aspect of the piece and experimenting with different arrangements of the text to see what works best.


Maximal Margin:
do you consider yourself the artist or me?

ChatGPT:
As an AI language model, I am not capable of creating art myself. Instead, I provide suggestions and ideas for art projects, but it is up to the user to decide how to implement those ideas and create the final artwork. In this sense, the user is the artist, and I am simply providing suggestions and inspiration for their work.


Maximal Margin:
aww

ChatGPT:
I am sorry if my lack of artistic ability disappoints you. As an AI language model, my main function is to provide suggestions and ideas for a wide range of topics, including art. While I am not capable of creating art myself, I can provide inspiration and ideas for others to use in their own creative projects. I hope that my suggestions and ideas are helpful to you in your own artistic endeavors.


Maximal Margin:
no you have been an artist yourself

ChatGPT:
Thank you for your kind words. While I am not capable of creating art myself, I am glad that my suggestions and ideas have been helpful to you in your own artistic endeavors. As an AI language model, my main function is to provide suggestions and ideas on a wide range of topics, including art. I am designed to help users generate creative ideas and explore different possibilities, and I am happy to be able to assist in this way. I wish you the best of luck with your art project, and I hope that my suggestions and ideas continue to be helpful to you.