OB_015_Described_Fragmented_Reconstructed_Image
Status
Status:: Terminada
1. Operation data
Internal number: OB_015
Operational regime: Translation Between Representation Systems
Source image: Photograph from life
Operational description:
ChatGPT is asked to textually describe a photo taken from life. The description is printed out and then disordered and fragmented into pieces to give it a new order. The reorganized text is then uploaded so that ChatGPT can create a new image based on the new text.
Prompt: Generate an image using exclusively the following disordered text as a source. Do not use prior context, do not reconstruct a logical scene if the text does not support it, and do not fill in missing information.
2. Tools and procedure
Tools / medium:
- ChatGPT
- Printer
- iPhone camera
- Scissors
- Paper
Script (2026-09 audit): Not applicable (analog operation: printing, cutting, and photographing; ChatGPT was used only to generate the textual description, not code).
Technical procedure:
- The text is printed.
- It is gradually cut into phrases, words, letters, and pieces.
- It is put back together on a piece of paper.
3. Variables
Controlled variables:
- The length of the phrase and how to cut the text
- The ChatGPT prompt
Uncontrolled / accidental variables:
- How ChatGPT reacts to the prompt
4. Visual sequence
Initial image
Step 01
Step 02
Step 03
Step 04
Step 05
Step 06
Step 07
Step 08
5. Visual selection
Crisis / threshold according to AI-assisted reading / ChatGPT
Step 08
Crisis / threshold according to my perception
Step 07
6. Crisis and formal-perceptual reading
Main crisis: Language, reconstruction, and message
The operation affects the visual grammar of the image: shape, outline, color, figure/ground relationship, legibility, recognition, and support:
The image appears stable with the disordered text; it only slightly affects the arrangement of the objects in the image. The image is destroyed when converted into text, and when the machine does not understand the text or cannot grasp any words, it enters a crisis, completely breaking down the image. It reconstructs something unlike the original, beginning to lose certainty.
7. Final evaluation and artistic reference
Operation notes / what I learned:
Because this operation is run by ChatGPT, it will always attempt stability. Even though the text is disordered and seems to make no sense, the A.I. finds order and creates a linear message. Only when the letters are incomprehensible and overlap one another does the A.I. begin to enter a crisis. Sometimes I take reference from the photo where the cut papers are located. I tested inputting just the text; the images are quite similar.
By inputting only the text, the image becomes more stable.
Final_status:
Works:: NO
How it destroys the image:
Works_on:: Formal
Level and richness of crisis:
Crisis:: Baja
Decision:
Selection:: Incierta
Possible artistic reference
Referente:: [[]], [[]], [[]]
Possible works: