NEW JERSEY · FACIAL RECOGNITION

Nijeer Parks spent ten days in jail while physical evidence that could have excluded him remained available to investigators.

False face match → 10 days jailed → civil challenge

Woodbridge police investigating a shoplifting case had fingerprints, DNA and other physical evidence connected to the actual suspect. Police nevertheless relied on a false facial-recognition result identifying Nijeer Parks. Parks was arrested in 2019 and spent ten days in jail.

The case is especially useful for understanding automation bias: surveillance did not eliminate other evidence. Instead, the machine-generated candidate allegedly became persuasive enough that investigators failed to wait for or adequately use evidence that could point elsewhere.

Why it matters: Human review does not automatically cure algorithmic error. A human can become anchored to the machine's answer and interpret the rest of the investigation around it.

Sources: ACLU of New Jersey case account · ACLU-NJ litigation record

WHY THIS LIBRARY EXISTS

This page preserves the anecdote and its caveats separately from the shorter Why Privacy Matters explainer. The goal is to keep the argument readable without hiding the receipts.