This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
Kimberlee Williams spent six months in jail after facial recognition technology incorrectly identified her as a suspect in Maryland bank fraud cases, despite her being in Oklahoma during the crimes.
In June 2021, Kimberlee Williams, a grandmother from Oklahoma, was arrested on Maryland warrants based on facial recognition technology that incorrectly identified her as a suspect in bank fraud cases from December 2019 and January 2020. A bank investigator had obtained security camera images of someone fraudulently withdrawing thousands of dollars from SunTrust and Truist branches in three Maryland counties. The investigator sent an image to Crimedex, a national listserv, where someone ran it through facial recognition technology and returned Williams' name as a match. Maryland police in Montgomery County, Prince George's County, and other jurisdictions relied solely on this facial recognition result and visual comparisons, making no attempt to verify Williams' alibi or investigate other leads. Williams was arrested at a military base checkpoint in Lawton, Oklahoma, and spent six months in jail - first in Oklahoma awaiting extradition, then in Maryland counties. She was eventually released when charges were dropped, but lost her job and continues to worry about future wrongful arrests. The report notes Williams is the fourteenth known person wrongfully arrested due to facial recognition failures, with most victims being Black individuals, though Williams and others show the technology affects people across racial groups.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
AI system
Due to a decision or action made by an AI system
Unintentional
Due to an unexpected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
