Key Takeaways
- Kimberlee Williams, an Oklahoma mother, was wrongfully jailed for six months in Maryland after police relied on a faulty facial‑recognition match and concealed that fact from the court.
- The arrest stemmed from a bank‑fraud investigation in which an unverified facial‑recognition search by an unknown private party flagged Williams as a suspect; no independent corroboration was performed.
- Montgomery County detectives obtained arrest warrants by misrepresenting the facial‑recognition lead as a visual identification, despite knowing the technology had produced a false match.
- Charges were eventually dropped in all three Maryland counties, but Williams spent 23 days in an Oklahoma jail and over three months in Maryland detention before release.
- The ACLU and ACLU of Maryland are demanding that Montgomery County, Prince George’s County, and Anne Arundel County police investigate the failures, issue a public apology, and adopt reforms to prevent similar injustices.
- Williams is the 14th publicly known case of wrongful arrest due to faulty facial‑recognition technology in the U.S.; the ACLU notes disproportionate error rates for people of color, women, older adults, and youths.
- Recommended policy changes include banning arrests based solely on facial‑recognition leads, prohibiting reliance on external, unverified facial‑recognition searches, and requiring independent corroboration before any warrant is issued.
Background of the Wrongful Arrest
In June 2021, Kimberlee Williams was accompanying her daughter on a DoorDash delivery to a military base in Lawton, Oklahoma, when base security discovered an outstanding Maryland arrest warrant for her. She was detained, transferred to Oklahoma authorities, and spent 23 days in a local jail before a Montgomery County officer arrived to transport her to Maryland.
How the Facial‑Recognition Match Occurred
The warrant originated from a bank‑fraud investigation in which an unknown suspect had impersonated account holders and withdrawn thousands of dollars from several Maryland bank branches. A bank investigator uploaded a surveillance image of the suspect to Crimedex, a national listserv for police and private investigators. An unidentified user ran the image through facial‑recognition software and returned Williams’ name and photograph as a purported match.
Police Reliance on the Unverified Lead
The bank investigator forwarded a memo to a Montgomery County detective stating that “facial recognition software” had flagged Williams as the suspect. Instead of treating this as a tentative lead, the detective sought an arrest warrant while concealing that the sole basis for the identification was the unverified facial‑recognition result.
Misrepresentation to the Magistrate
When applying for the warrant, the detective told the magistrate judge that Williams had been “identified” as the suspect and that he had visually compared a photo of the suspect with an older photo of Williams to confirm the match. In reality, the visual comparison was meaningless because the facial‑recognition system had returned an innocent person who merely resembled the suspect—a known limitation of the technology.
Initial Detention and Transfer
After the warrant was issued, Williams was held in an Oklahoma jail for 23 days. Upon arrival in Maryland, she was placed in Montgomery County detention, where she remained for more than three months while contesting the charges. She steadfastly maintained she had never set foot in Maryland before her arrest.
Charges in Additional Counties
While Montgomery County prosecutors ultimately dismissed the case in October 2021, similar charges remained pending in Prince George’s County and Anne Arundel County, all based on the same flawed facial‑recognition match and lacking any independent investigation. Williams was transferred between jails in those counties, spending another two months fighting the allegations before all charges were dropped in December 2021.
Impact on Williams and Her Family
Williams lost six months of her life—time she could not recover—while being separated from her children, her home, and her job. She emphasized that she had never been to Maryland prior to her wrongful arrest and described the ordeal as a devastating intrusion caused by an algorithmic error. Her experience underscores the human cost of relying on unchecked facial‑recognition technology.
Broader Pattern of Misidentifications
Williams is the fourteenth person publicly known to have been wrongfully arrested because of faulty facial‑recognition matches in the United States. The ACLU notes that such errors disproportionately affect people of color, women, older adults, and youths. Similar mistaken arrests have been documented in police departments across Michigan, Missouri, Louisiana, Nevada, New Jersey, North Dakota, Florida, and Arizona, among others.
ACLU’s Response and Settlement Precedent
In 2024, the ACLU secured a landmark settlement with the City of Detroit following the wrongful arrest of Robert Williams, another victim of facial‑recognition misidentification. The settlement highlighted the need for accountability and reform. The ACLUnow uses Williams’ case to press Maryland police departments for similar redress.
Calls for Investigation and Apology
The ACLU and ACLU of Maryland have sent formal letters to the police chiefs of Montgomery County, Prince George’s County, and Anne Arundel County demanding that they: (1) conduct a thorough internal investigation into the failures that led to Williams’ arrest; (2) issue a public apology acknowledging the harm caused; and (3) implement concrete policy changes to prevent recurrence.
Recommended Policy Reforms
To curb future injustices, the ACLU advocates that police departments:
- Prohibit reliance on facial‑recognition leads obtained from outside entities without independent verification.
- Ban arrests based solely on a facial‑recognition match followed by a superficial visual comparison, which is inherently tainted when the technology misidentifies an innocent look‑alike.
- Require corroborating evidence—such as eyewitness testimony, physical evidence, or additional investigative work—before any arrest warrant is sought.
- Mandate training on the known limitations and bias risks of facial‑recognition systems, especially regarding demographic groups that experience higher false‑match rates.
Conclusion
Kimberlee Williams’ six‑month ordeal illustrates the grave dangers posed when law‑enforcement agencies treat unvetted facial‑recognition outputs as definitive proof of guilt. Her case adds to a growing list of wrongful arrests tied to this technology and underscores the urgent need for transparency, accountability, and robust safeguards. By investigating the missteps, apologizing to the victim, and enacting the recommended reforms, Maryland police departments can begin to rectify this injustice and help ensure that no other family suffers a similar loss of liberty.