[ISEA97] Artist Talk: Christopher P. Csikszentmihalyi – Character Input

Artist Statement

The Character Input fiction is a neural network which can read aspects of a person’s character from an image of their face. The phrenological tool was developed by an Estonian scientist who continued to work with perceptrons and neural networks through the 60s and 70s, when they had fallen out of popularity in the west. His initial research into the reading of gender from photographs of subjects eventually led to an amazing network which could read the most personal and specific aspects of the subject’s personality.

The piece alludes the cultural agenda around measuring human attributes: from IQ tests to the Human genome project. Contemporary cognitive scientists are actually using neural networks to read facial emotion; this and related technologies seem to owe a strong debt to phrenology and cranioscopy. The Character input installation consists of a large image bank of faces, a variety of old communist-block computers and interfaces, and an interactive section which allows the viewer to be scanned by the neural network, which then reveals aspects of their personality. In addition, a Soviet-style slide-show documentary describes the work of the scientist, interleaved with relevant events (workers celebrate 3rd year of 5 year plan, etc.).

  • Christopher P. Csikszentmihalyi (U.S.A.), Visual Arts, University of California, San Diego, USA.   edgyproduct.org