And another one ... Detroit remains epicenter of false arrests from facial rec - DINT 143
Detroit police say facial recognition didn't lead to LaDonna Crutchfield's false arrest. Redacted court documents paint a different picture.
Detroit’s third documented case of tech-based false arrest reveals a pattern of missteps
They handcuffed her after she slipped walking to the police vehicle.
Her kids watched, as did her niece who was living with her at the time.
They were arresting her on outstanding warrants. She had to go to court to appear before a judge.
All of this seems reasonable until you realize it’s a case of mistaken identity and LaDonna Crutchfield was on her way to missing her shift at her second job; all because of facial recognition.
This scene unfolded in the same place the last two major mistaken identity / facial recognition arrests happened: Detroit.
The city has a track record of arresting working-class Black people under the pretense photo identification.
First, there was Robert Williams, a father of two who as nowhere near the scene of a crime police arrested him for in 2020. Four years later the city paid Williams $300,000 and agreed to federal supervision of their facial recognition practices.
Then, there was Porcha Woodruff, a pregnant women getting her kids ready for school who was accused of a carjacking. Woodruff was eight months pregnant at the time, making her an unlikely suspect. Woodruff’s case, led by attorney Ivan Land, is still making its way through the court system.
And the latest name we can add to Detroit’s list of false arrests? LaDonna Crutchfield; a woman who was taking a break between her two jobs, reading a book to her five-year-old daughter in their home, when the police knocked. They told her to come with them and to calm down when she asked what it was all about. A few hours later Crutchfield had been released by the authorities, making it to her job with mere minutes to spare.
Note: Crutchfield had to surrender DNA samples before the police would release her ina timely manner. She had a tough choice, retain her biometric data which could be manipulated against her or jeopardize the income she earned from her second job.
Not that she was able to be productive at her second job. She couldn’t stop crying. Handcuffed, falsely accused, marched to a squad car in front of her children and neighbors. Almost losing her job.
Last week, Crutchfield had Ivan Land file a lawsuit naming the City of Detroit and the arresting officers as plaintiffs.
While there are several disturbing facts about the case, a few stand out as more poignant than others.
The city claims the arrest didn’t stem from misuse of facial recognition systems.
But a review of the court documents reveals a heavily redacted document which hides just how the city acquired its evidence to arrest Crutchfield.
Here are results from an analysis tool I used to unearth the truth about the redacted court document:
Analysis of Facial Recognition Evidence
The Case Supplemental Report (Exhibit 7) doesn't explicitly use the term "facial recognition," but provides strong circumstantial evidence that this technology was employed:
Database Search Reference: The report states "On December 29, 2023 4:00 pm, I conducted a database search..." followed by redacted text, then continues with "Ms. Crutchfield matched the suspected I observed through greenlight with the shooting weapon."
Project Green Light Connection: The term "greenlight" refers to Project Green Light Detroit, a public-private partnership where businesses install high-definition cameras that feed directly to the Detroit Police Department's Real Time Crime Center. This system is known to be integrated with facial recognition capabilities.
Timing and Process: The report shows the database search occurred on December 29, but Ms. Crutchfield wasn't arrested until January 23, with no documentation of any additional investigation during that nearly month-long gap.
Pattern of Practice: This matches the pattern described in the complaint regarding the Williams and Woodruff cases, where facial recognition was used to identify suspects without sufficient additional investigation.
Strategic Redactions: The extensive redactions in the report likely cover technical details about the identification process, possibly including specific reference to facial recognition technology.
Significance of the redacted document to the case
This document is legally significant because:
It establishes the causal link between the database search (likely facial recognition) and Ms. Crutchfield's identification as a suspect.
It shows the identification was based primarily on comparing Green Light footage to a database rather than traditional investigative techniques.
It contradicts any defense claim that extensive investigation occurred before arrest.
When combined with the physical dissimilarities between Ms. Crutchfield (5'07", 265 lbs) and the original suspect description (6'0", 300 lbs), it suggests that over-reliance on potentially flawed technology led to the mistaken arrest.
The absence of explicit "facial recognition" terminology in the document isn't unusual - police departments often use general terms like "database search" or "image comparison" in official reports even when facial recognition is the specific technology employed.
What conclusions can we draw from these cases? Detroit has a clear pattern of low-effort policework when using facial recognition technology to solve crimes. Previous interventions haven’t worked to improve the quality of policing through technology in Detroit. Black people are disproportionately affected by this lack of focus, professionalism, and attention to detail.
–
Here are the background documents and articles used to research this case:
February 25, 2025 Detroit woman suing police, claiming faulty facial recognition technology led to unjust arrest
July 11, 2024 Detroit Police Department and ACLU reach agreement over use of facial recognition technology
More News at the Intersection of Tech, Race, and Gender
First Thought … With Google’s track record with AI and people of color, we’re not sure this will help us
Palate Cleanser
