Detroit police facial recognition misidentified a pregnant woman, causing a wrongful arrest
On February 16, 2023, Detroit police arrested Porcha Woodruff, who was eight months pregnant, after DataWorks Plus facial recognition software matched her to surveillance footage of a carjacking and robbery suspect. She was held for approximately 11 hours at the Detroit Detention Center before being released on a $100,000 personal bond, and the criminal case was dismissed on March 6, 2023 for insufficient evidence. Woodruff filed a federal civil rights lawsuit in August 2023, which was dismissed in September 2025 after the judge ruled the detective had probable cause at the time of the arrest.
A facial recognition algorithm produced a false match on an eight-months-pregnant woman, and investigators failed to notice the suspect in the surveillance video was visibly pregnant before making the arrest.
Key facts
- What
- On February 16, 2023, Detroit police arrested Porcha Woodruff, who was eight months pregnant, after DataWorks Plus facial recognition software matched her to surveillance footage of a carjacking and robbery suspect.
- Incident date
- Feb 16, 2023
- Who
- Detroit Police Department
- Failure mode
- Hallucination
- AI surface
- Search / RAG
- Severity
- High
What happened
On February 16, 2023, Detroit police officers arrived at Porcha Woodruff's home while she was getting her two children ready for school and arrested her on charges of robbery and carjacking based on a facial recognition match from DataWorks Plus software. Detective LaShauntia Oliver had run facial recognition on gas station surveillance video and then included an eight-year-old booking photo of Woodruff in a photo lineup shown to the victim, who identified her. Woodruff was held for approximately 11 hours at the Detroit Detention Center and released on a $100,000 personal bond; after her release she sought medical treatment for dehydration, a low heart rate, and stress-related contractions. The Wayne County Prosecutor's Office dismissed the criminal case on March 6, 2023 citing insufficient evidence, and Detroit later changed its policy to prohibit arrests based solely on facial recognition results or photo lineups generated from facial recognition searches.
What broke inside the model
- 01 · TriggerA user asks for a fact, a citation, or a figure.
- 02 · Model stepThe model writes a fluent, confident answer.
- 03 · Control gapNothing ties the claim back to a real source.
- 04 · FailureA fabricated fact ships as if it were verified.
- 05 · ConsequenceThe false claim reaches a customer, a court, or the public.
Confidence holds, and even spikes, as the claim detaches from any source.
The DataWorks Plus facial recognition system returned a false positive match between Woodruff and a suspect captured on gas station surveillance video, an error compounded by the investigating detective's decision to include an eight-year-old booking photo of Woodruff in a photo lineup shown to the victim instead of her current driver's license. Facial recognition algorithms are known to perform less accurately on women and people of color, and the system produced a confident but incorrect match that was treated as sufficient investigative evidence without corroboration. The detective's police report also failed to note that the suspect in the video appeared pregnant, which would have been an obvious disqualifying detail.
What it cost
Sources
- PressDetroit woman sues city after being falsely arrested while pregnant due to facial recognition technologynbcnews.com
- PressWoman wrongly accused of carjacking loses lawsuit against Detroit police who used facial techapnews.com
- Court FilingWoodruff v. Detroit, City of, 5:23-cv-11886courtlistener.com
Cite this entry
https://failureindex.ai/failures/detroit-police-facial-recognitionAI Failure Index. "Detroit police facial recognition misidentified a pregnant woman, causing a wrongful arrest" (FI-0110). Realm Labs. https://failureindex.ai/failures/detroit-police-facial-recognition (indexed Jun 4, 2026).Data fields CC-BY 4.0, prose citation permitted. Incident ID FI-0110. Full dataset at /data.
Note from Realm Labs, the Index steward
How Realm would have caught this
- Prism
- OmniGuard
- AI Detection & Response (AIDR)
A runtime layer that watches the model's internal state can flag the moment a model commits to a claim it has no support for, and hold or reroute the response before it reaches a user. Realm reads those signals in real time rather than grading the transcript after the fact.