1
There is an image.
The image was generated in approximately forty seconds. It shows a knight in a forest. The knight is mounted. The forest is dark. The light is the light of late afternoon, or of dawn, or of a particular quality of digital rendering that resembles both. The knight is facing something.
It is not clear what.
2
The image was produced by a text prompt. The prompt is not known.
The model that received the prompt had been trained on images.
Among the images it had been trained on were photographs of armor in museum cases. Photographs of helmets behind glass.
Photographs of objects that had been made by hand by people who had spent years learning to make them, conserved by people who had spent years learning to conserve them, cataloged by people who had spent years learning the periods and the provenance and the metallurgy required to describe them, housed in departments that had taken decades to establish and require ongoing endowment to maintain.
The model compressed these into weights.
The weights produced the knight. The weights produced the forest.
3
Consider the armor in the image.
The armor is of no specific period. The helmet resembles a sallet, which was common in the mid-fifteenth century in Germany and Italy. The breastplate resembles a cuirass of the sixteenth century. The gauntlets are from neither period in particular.
Together they form armor that was never worn, in a century that never existed, by a knight who was never born.
The actual sallet in the collection of the Metropolitan Museum of Art, accession number 29.158.27, acquired in 1929, was made in southern Germany around 1480. Its surface shows marks from a hammer.
Its interior shows marks from a head. It weighs approximately 2.5 kilograms. It was worn. The name of the person who wore it is not recorded. The name of the person who made it is not recorded.
The image has no weight. None of the rest of this is in it either.
4
Consider the forest.
The forest in the image is a forest assembled from the visual residue of other forests: the forests of nineteenth-century German Romantic painting, the forests of fantasy illustration, the forests of films set in the Middle Ages, the forests of other generated images of knights in forests. It is a forest made of forests made of forests.
It has no particular trees. It is not an oak forest or a beech forest or a pine forest. It is a forest in general, which does not exist.
There is a forest that exists.
Jan Długosz described it in his Annals in the 1460s, in Latin, in the forests of northern Poland, in a manuscript copied by hand and now held in the Czartoryski Library in Kraków.
That forest has particular trees. Some of those trees are still there.
Bialowieza forest
(photo credit) https://www.earth.com/news/fate-of-primeval-bialowieza-forest-hangs-in-the-balance/
5
In 1867, a boy was born in New York City. His name was Bashford Dean. When he was six years old, visiting the home of a family acquaintance in Yonkers, he spent several hours examining a helmet from the man’s collection of Asian and medieval arms and weaponry. He did not stop thinking about helmets.
He completed a PhD in zoology and paleontology at Columbia University in 1891. His research took him to Europe, Russia, Alaska, Japan, and the Pacific coast of the United States. He became a full professor at Columbia and the founding curator of the Department of Reptiles and Fishes at the American Museum of Natural History.
He is the only person ever to have held concurrent curatorial positions at the American Museum of Natural History and the Metropolitan Museum of Art.
In 1904, he began building the Metropolitan’s collection of arms and armor.
In 1912, the Department of Arms and Armor was formally established, largely because of him. He spent the following decades acquiring, cataloging, and studying objects from across Europe and Asia, including Japanese armor, of which he had assembled the most significant collection outside Japan.
During the First World War, the United States government commissioned him to design helmets for American soldiers.
The research that had started with a six-year-old standing in front of a helmet in Yonkers informed American helmet design through at least the 1980s. The M1 helmet, worn by American soldiers in the Second World War, descended from his work.
He died in 1928, one year before the Metropolitan acquired the sallet.
6
The hammer marks. The weight. The name of the man who wore it. The name of the man who made it. The forests of northern Poland. The particular trees. The Latin of Długosz. The six-year-old in Yonkers who could not stop looking. Columbia University, 1891. Russia. Alaska. Japan. The M1 helmet. The men who wore it.
The men whose lives were saved because of it.
Etc.
7
The AI-generated image circulates. It takes less than a second to share the image. It is shared. It is shared again.
The sallet, accession number 29.158.27, does not circulate. It stays in its case.
It was worn once. The marks are still there.
#ArtificialIntelligence #AIArt #CulturalHeritage #DigitalCulture #Creativity
