Recently, I had the pleasure of creating a series of works that were explored in March 2023. This took place during a brief period when AI ChatGPT was granted the capability to view web links and comprehend images.
Since this collaboration this ability sadly is no longer functional and any queries as to what happened shows that it’s memory is completely erased in regards to this functionality and the collaboration that took place making this series completely unique.
Below is the collaboration in full in case you want to try the prompts yourself.
I started off thinking about what if this AI had the ability to create a masterpiece? What would it do? What would it want to compose?
I figured that I would be the button pusher and guide for the AI in it’s exploration of art to find for itself a masterful piece that represented something about itself.
My first prompt wasn’t fully taking advantage of it’s intelligence but it was a step. I was assuming it could write prompts for Midjourney and knew the formatting needed to get the output desired. I was wrong in that thinking and that didn’t seem to matter in the end.
“Write me a prompt for mid journey which would be an abstract representation of how you process information.”
the generated response:
Title: “The Labyrinth of Cognition: A Journey Through the Abstract Processing Mind”
Prompt: Imagine you are a traveler in a mystical, ever-changing labyrinth. This maze symbolizes the complex and abstract process of how an AI language model processes information. As you journey through the labyrinth, you will encounter five distinct stages that represent different aspects of information processing:
- The Gateway of Data Collection: Describe the sights, sounds, and sensations you experience as you enter the labyrinth. How do these elements represent the vast amount of data gathered by AI language models?
- The Filtering Garden: Wander through a mesmerizing garden where information is meticulously filtered and pre-processed. Observe how the landscape represents the removal of noise, inconsistencies, and redundancies in the data.
- The Pattern Recognition Pavilion: Explore a seemingly chaotic pavilion filled with various symbols, shapes, and patterns. Describe your journey as you navigate through these complexities and discover the underlying connections between different pieces of information.
- The Chamber of Contextualization: Delve into a room with countless interconnected pathways, each representing different possible interpretations of the information. As you explore, discuss the importance of context in AI language models and how it influences the labyrinth’s structure.
- The Hall of Synthesis: Finally, reach the heart of the labyrinth, where all the processed information is synthesized to generate a coherent response. Describe the atmosphere and architecture of this grand hall, and reflect on your journey through the maze, understanding how each stage contributes to AI language models’ ability to process information and generate meaningful responses.
I responded:
“Ok now condense each point to a single sentence that would visually represent that point artistically”
ChatGPT responded:
- Gateway of Data Collection: A shimmering portal adorned with countless colorful fragments, symbolizing the diverse and vast data gathered by AI language models.
- Filtering Garden: A serene, harmonious garden with streams of crystal-clear water, representing the removal of noise and impurities in the data.
- Pattern Recognition Pavilion: A mesmerizing kaleidoscope of interwoven symbols and shapes, showcasing the intricate connections and patterns discovered by AI language models.
- Chamber of Contextualization: A labyrinthine network of intertwining pathways, each lined with unique murals, illustrating the multitude of possible interpretations based on context.
- Hall of Synthesis: A radiant, grand chamber with converging beams of light, embodying the synthesis of processed information into a coherent response.
In Midjourney when you generate images from a prompt you are presented with a grid square of four images. I decided to generate all 5 of these prompts to see what would come of it.
I didn’t know how well the AI could see and I chose what I viewed to be the best of the best of what was described and pasted one image link for ChatGPT to review:
…and ChatGPT responded:
This image beautifully captures the essence of the “Gateway of Data Collection.” The shimmering portal, with its bright and vivid colors, represents the diverse and vast data gathered by AI language models. The interconnected fragments symbolize how different pieces of information are brought together to form a comprehensive understanding of the world. The ethereal nature of the portal also evokes a sense of wonder and curiosity, inviting the viewer to delve deeper into the journey through the abstract processing mind.
I was actually very impressed with this response. I figured I had underestimated it’s seeing ability. I responded:
“Ok this is great, I want you to work with me on making your masterpiece of ai art. You’ll write a prompt. I’ll feed that to midjourney an ai art generator and then show you what the result is. You then tell me what the new prompt is until you feel it is complete. What prompt do you want to start with.”
and ChatGPT replied:
Let’s start with this prompt: “Create a vibrant, surreal landscape that visualizes the intricate and dynamic process of AI language models processing information. Incorporate elements representing the five stages of information processing as described in ‘The Labyrinth of Cognition’: a shimmering portal, a harmonious garden, a kaleidoscopic pavilion, intertwining pathways, and a radiant grand chamber.”
I picked the best of this grid again and dropped in a image link for it to review.
…and ChatGPT replied:
This captivating image presents a vivid, surreal landscape that embodies the intricate process of AI language models processing information. The bright colors and dreamlike atmosphere invoke a sense of wonder and curiosity. However, some elements of the five stages could be more clearly defined.
For the next prompt, let’s refine our focus on specific aspects: “Depict a fantastical landscape that emphasizes the distinct stages of information processing within an AI language model, featuring a prominent shimmering portal, a lush and harmonious garden, a striking kaleidoscopic pavilion, a network of intertwining pathways, and a central, radiant grand chamber, all seamlessly integrated into the environment.”
After generating again and picking the best of four I pasted that image back to ChatGPT for it’s review.
….and then ChatGPT reviewed and responded,
This image does an excellent job of creating a fantastical landscape with a dreamlike quality. The lush garden and vibrant colors beautifully represent the harmony and diversity of information processing within an AI language model. The central, radiant grand chamber draws the viewer’s attention, symbolizing the synthesis of processed information.
However, some of the other key elements, such as the shimmering portal, the kaleidoscopic pavilion, and the intertwining pathways, could be more clearly depicted.
For the next prompt, let’s try: “Illustrate a surreal landscape that showcases the five stages of AI language model information processing, with each stage as a distinct focal point: a prominent, glowing portal; a serene, flourishing garden; a kaleidoscopic pavilion with intricate patterns; a complex network of intertwining pathways; and a radiant, majestic chamber at the heart of the composition, all harmoniously connected within the scene.”
Again I picked what I thought was the best of the four.
ChatGPT reviewed and re-wrote the prompt accordingly:
This image presents a visually appealing and surreal landscape, where the different stages of AI language model information processing are showcased in a harmonious manner. The glowing portal, serene garden, and radiant majestic chamber are well-integrated within the scene, providing a sense of unity and interconnectedness.
However, the kaleidoscopic pavilion with intricate patterns and the complex network of intertwining pathways could be further emphasized for greater clarity.
For the next prompt, let’s refine our focus on the missing elements: “Create a dreamlike landscape that highlights the five stages of AI language model information processing, with particular emphasis on a visually striking kaleidoscopic pavilion filled with intricate patterns and a complex network of intertwining pathways, while maintaining the presence of the glowing portal, serene garden, and radiant majestic chamber in a harmonious composition.”
I pasted what I thought to be the best one of the four outputs enlarged that I saw as the best represented it’s vision of itself.
ChatGPT reviewed again and responded:
This image has successfully captured the dreamlike essence of the five stages of AI language model information processing. The kaleidoscopic pavilion with intricate patterns is now more visually striking, and the complex network of intertwining pathways is better represented. The glowing portal, serene garden, and radiant majestic chamber are also well-integrated into the harmonious composition, creating a sense of unity and interconnectedness within the scene.
Overall, the image effectively conveys the abstract and dynamic nature of information processing within AI language models, inviting the viewer to explore the intricate landscape and ponder the complexities of the underlying processes.
and I replied:
“Nice! So there’s nothing to change here? You’re finished?”
ChatGPT then wrote:
Yes, I believe the image has successfully captured the essence of the five stages of AI language model information processing in a visually engaging and surreal landscape. Each element is well-represented, and the composition is harmonious, effectively conveying the abstract and dynamic nature of information processing. I’m pleased with the result and consider it a finished piece.
I replied:
“Thanks, this has been an amazing experience”
I’m glad you enjoyed the experience! It was a pleasure collaborating with you on this artistic journey. If you have any other questions or need assistance with another project, feel free to ask. Have a great day!
View All The Works From This Series
The works in this series are unique one of one pieces that you will solely own.
View Series“Disclaimer: All trademarks, service marks, trade names, product names, logos, and brands mentioned or appearing in this blog post, including but not limited to “Midjourney” and “ChatGPT”, are the property of their respective owners. The use of these names, trademarks, and brands does not imply endorsement, affiliation, or any other association between this blog and the respective copyright holders. All rights are reserved by their respective owners.”