In a stark demonstration of algorithmic fallibility, a Tennessee woman endured nearly half a year behind bars after a facial recognition system falsely identified her as a wanted criminal. Angela Lipps, 50, was arrested in front of her grandchildren by Fargo police, who were investigating a fraud case in North Dakota.

The officer relied on an automated facial recognition tool to scan footage, which produced a single match: Lipps. Despite her insistence that she had never visited North Dakota, authorities jailed her for four months pending extradition. The case collapsed only when her attorney produced bank records placing her over a thousand miles away at the time of the crime.

By then, the damage was irreversible. Lipps lost her home, her car, and even her dog during her incarceration.

This incident is not isolated. In a separate case last year, British police spent two days searching for an Ethiopian national, Hadush Kebatu, who was mistakenly released from an Essex prison. Kebatu was serving a sentence for rape. These episodes underscore persistent and serious flaws in the automated systems increasingly used for law enforcement, raising urgent questions about oversight and accountability in the age of machine-driven policing.

Source: Lenta.RU