XSci

Artificial Intelligence in Film Postproduction: A Critical Review of Selected Applications

Nicholas Tanner

Submitted by Nicholas Tanner

Published October 9, 2026 · Version v7, October 22, 2026 · DOI 10.66977/xsci.2610.000t

Film Studies, Artificial Intelligence

Abstract

This interpretive critical review examines the evidence for judging the suitability, creative value and ethical use of artificial intelligence (AI) in film postproduction. It analyses 14 empirical or qualitative studies and ten groups of related production reports, alongside scholarship on editing and documentary ethics. The central question is what preserving a performance requires when AI alters appearance or speech and editing shapes the resulting portrayal. Accounts of Here and Welcome to Chechnya illustrate different aims: retaining an actor's contribution while changing apparent age, and retaining expressive motion while disguising identity. These accounts describe filmmakers' intentions without demonstrating creative or audience effects. The analysis distinguishes correction to meet an agreed requirement from creative development, while recognising that both can occur in the same operation. The time required for this work cannot, by itself, establish a creative benefit or a repair burden. Evaluating an altered image or voice also differs from evaluating the scene in which it appears. Dubbing studies show that measures of emotion, reconstruction quality, speech recognition and synchronisation can rank systems differently. An earlier judgment also does not automatically justify a later use or establish that an alteration protects a documentary subject's identity. The review argues that evaluation should specify the intended change, the contribution to preserve and the portrayal produced in the finished scene. Purposive selection limits the conclusions to the selected material. Inferences affected by six unresolved methodological and reporting ambiguities are withheld; proposed evaluations and safeguards remain untested.

View PDF