ATLAS OF ARTIFICIAL ILLNESS

BACKGROUND

Artificial intelligence (AI) and machine/deep learning (ML) are beginning to revolutionize modern medical practice. These modalities represent a powerful diagnostic tool for physicians that can elucidate patterns of disease with staggering accuracy, particularly with image analysis. Images of patient pathology can be presented to AI/ML systems, returning accurate diagnoses validated by physicians.

What about attempting the opposite? What is the image produced by AI/ML when a diagnosis is used as the input for an algorithm? What is the non-human visual expression of the wide spectrum of human pathology, from psychiatric conditions to musculoskeletal disorders?

This project explores the human spectrum of disease, as viewed from the lens of machine learning and artificial intelligence.

PROJECT DETAILS

The Atlas of Artificial Intelligence is created by presenting prompts to machine learning image creation software in the following systematic fashion:

• Each image prompt was based on a singular pathology/illness, including the diagnosis, cardinal patient symptoms, and corresponding ICD-10 code, followed by various modifiers

• Four images were produced from the initial prompt. One image was selected RANDOMLY. This image was then re-rolled at least 5 times - finally the most compelling image was selected

ROADMAP/UTILITY

Multiple volumes of the Atlas will be released.

Volume 1: Phobia

Volume 2: General Pathology

Volume 3: TBD

20% of all project earnings will be donated to the National Institute on Deafness and Other Communication Disorders (NIDCD)

VOLUME 1 - PHOBIA

• 100 tokens, ERC-721 / ETH

• All pieces are 1/1

VOLUME 2 - GENERAL PATHOLOGY

• 300 tokens, ERC-721 / ETH

• All pieces are 1/1