86 min | NR | November 11, 2020 | 7th Empire Media
An MIT researcher discovers that facial-recognition software cannot see her dark-skinned face until she puts on a white mask. From that one glitch, the film builds a case against the algorithms already deciding who gets a job, a loan, and a place to live. The machines are not neutral. They learned from us.
Joy Buolamwini, a researcher at the MIT Media Lab, builds an art project and finds that the facial-recognition system cannot detect her until she holds a white mask over her face. That glitch becomes the spine of the film. Shalini Kantayya uses Buolamwini’s discovery as a doorway into the larger machinery of algorithmic decision-making. The film is not really about facial recognition. It is about who writes the code, whose data trains it, and which people get classified as threats by systems that claim to be objective.
Buolamwini carries the film as its central figure and its emotional anchor. She testifies before Congress and explains bias in math with the clarity of someone who has done the work and cannot believe she has to argue the point. Cathy O’Neil, the mathematician who wrote about weapons of math destruction, speaks with the bluntness of a quant who left finance and saw the damage from the inside. Meredith Broussard, Safiya Umoja Noble, and Virginia Eubanks each press on a different failure, from the myth of technological neutrality to the way automated systems punish the poor first. Silkie Carlo takes the film to London and confronts police using live facial recognition on the street, and her arguments with officers turn abstract policy into a person being stopped on a sidewalk.
Kantayya directs and writes the film as advocacy journalism, and she structures it around faces and the failure to read them correctly. The camera lingers on the detection boxes that snap onto white faces and slide off Black ones, and that visual becomes the film’s recurring argument. She cuts between the United States and China to show two versions of the same surveillance logic. The editing favors clarity over complexity, moving the viewer through technical material without losing the thread. The film chooses accessibility as its method, and it pays for that choice by keeping the formal ambition low.
This is a primer, and it knows it is a primer. Kantayya wants the audience to leave understanding that an algorithm is an opinion embedded in code, and she delivers that lesson with discipline and force. The film is stronger as a brief than as cinema, and it trades depth on any single case for breadth across many. What it captures is the moment a researcher realizes the machine cannot see her, and decides to make everyone else see the machine.