THE HUMAN RECORD

Human beings have always been turned into records: photographs, names, forms, classifications, files, prices and databases. Records allow people to be preserved, compared, administered and understood, but they also reduce a living person into a form that a system can hold.

THE HUMAN RECORD began from an interest in the information asymmetry between a person and the systems that claim to describe, classify or value them. A photograph records a surface. An archive preserves what remains. An institution establishes categories. A market assigns value. A computational system infers patterns from limited information. Each produces knowledge about a person, but none possesses the person in full.

Information asymmetry is also temporal and spatial. In markets, institutions and everyday interactions, outcomes are not determined by accuracy alone. They are shaped by who is able to define the situation first, occupy attention, establish a category, and have that description recognised as authoritative.

Legal and institutional authority can determine what is compliant, enforceable or officially recorded, but authority does not automatically make a representation true, complete or morally just. A record may therefore become consequential before it becomes accurate.

Artificial intelligence makes this condition especially visible. A machine can generate a confident description from incomplete material, yet confidence does not make that description neutral, complete or final.

The central question is therefore not whether AI can produce an accurate portrait. It is who is granted the authority to describe a person, and whether the person being described still retains the right to answer, correct, reject or remain unresolved.

Alongside contemporary participant records, BEFORE THE MODEL introduces found instant photographs made before today's machine-readable systems. Many of the people within these photographs can no longer explain the circumstances in which the images were made or respond to the meanings later viewers place upon them.

The record survives while the subject's capacity to answer disappears.

This suggests that incomplete representation did not begin with artificial intelligence. AI intensifies a much older condition of photography and the archive.

The use of discontinued peel-apart film is part of this tension. Its chemical residue, physical instability and finite materiality resist the frictionless appearance of digital systems.

The photograph exists as an object that ages, stains, separates and remains materially incomplete. It does not behave like a perfectly reproducible unit of data.

For this reason, THE HUMAN RECORD does not seek a final or definitive portrait.

Photography, handwriting, voice, self-description, machine-generated interpretation, participant response and visual reconstruction are preserved beside one another as competing forms of record.

Disagreement is not corrected out of the archive; it is kept visible.

The person being recorded is allowed to answer the system rather than simply becoming its output.

The project therefore treats incompleteness not as a failure to be solved, but as a condition that protects human agency.

A HUMAN MAY ENTER THE RECORD,
BUT NO RECORD HAS FINAL AUTHORITY OVER THE HUMAN.