Hey, sorry about this.
James Hopper
12

You are assuming that cultural context is somehow an impregnable input into an artificial intelligence.

Parsing cultural content and finding explanatory variables seems right up the alley of a deep brain. The issue is how to normalize variables into a coherent set of data (video, Twitter, people magazine, CNN, photos), but neural networks keep solving this problem at a local level. When we have the infrastructure that can fluidly use and switch between different agents for different purposes, understanding, categorizing and bringing back a human cultural context is trivial. And this can be abstracted further to the communication patterns of any wetware.

Therefore, when we narrow the role of the human to choice and curation of machine-created imagery, because what used to be visual art is now instant illustration, it is meaningful to highlight that a narrowing is actually occurring. It is remarkable! But it does not imply that there is no place for humans in software art.

Imagine the following.

A person sits on a couch wearing a VR headset, or looks at a wall that functions like a monitor. The person says out loud: “Echo, create art for me.”

The sound is converted to a command that wakes up an artificial artist nexus. It connects with our person’s Facebook, Twitter and Google Photos graph. Based on the outliers in photo preferences (cat or dog person, children or not, loner or partied), analysis of language patterns (personality mapping), and social data to place this person in context, the artificial artist selects a style which creates the highest emotional reaction on this particular individual. This behavior is learned by the artist over millions of practice sessions, reading human faces for emotional expression. It then decides on a recursive 3d fractal that serves as the landscape for a particular expression, with or without human figures and simulated narrative, and renders the result in real time in projected or virtual reality. Random error with a bias for human preference is introduced to create the right amount of innovation and challenge (for this individual) based on prior experience.

Yes, the person said “Create art for me.” But who is the artist? The mathematician that taught the neural network how to read normalized statistical data? The librarian that uploaded human art styles into the algorithm? The coder that combined these agents together? Or the software that had learned in thousands of ways on millions of examples?