Robert Julian-Borchak Williams was arrested in January after Detroit police told him to turn himself in, according to The New York Times. Facial recognition software had identified him as the person seen in surveillance footage of a 2018 shoplifting. Williams spent 30 hours in prison before posting bail, and the case was later dismissed.
The software Detroit police used displayed a warning alongside its results: “This document is not a positive identification. It is an investigative lead only and is not probable cause for arrest.” Despite this, critics say the technology was the sole basis for Williams’s arrest.
Overreliance on Flawed AI
Facial recognition developers have repeatedly warned against using their technology as the only evidence for an arrest, especially when working from grainy surveillance images. The mismatch in Williams’s case was clear, yet he still faced arrest and detention.
Clare Garvie, a lawyer at Georgetown University, told The New York Times: “I strongly suspect this is not the first case to misidentify someone to arrest them for a crime they didn’t commit. This is just the first time we know about it.”
Williams’s experience points to a broader problem of police overusing unreliable algorithms. While he was released and the case dismissed, the incident raises questions about how often such errors occur without public knowledge.
Editorial Notes