This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
In 2019, [name withheld] was wrongfully arrested and spent nearly four years in pretrial detention after facial recognition software identified him as a suspect in an armed robbery. Despite having an alibi, he was charged based on the identification, which lacked transparency and was later challenged in court. The case became a landmark for criminal justice reform, highlighting the risks of unchecked surveillance power and the lack of accountability frameworks for AI tools used by law enforcement.
In 2019, facial recognition technology misidentified Francisco Arteaga as a suspect in an armed robbery in New Jersey. The incident led to nearly four years of pretrial incarceration. Despite having an alibi, Arteaga was charged based on the flawed identification. The legal battle that followed resulted in a court ruling requiring police to reveal details about the algorithms used in facial recognition. The process exposed significant gaps in transparency and accountability.
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.
Also classified as
Law enforcement officers over-relied on a single 'possible match' from the AI system to arrest and charge the suspect despite a lack of supporting evidence.
The proprietary nature of the AI software and lack of cross-jurisdictional transparency prevented the defendant from accessing the system's source code and error rates for his defense.
The AI system failed to perform reliably by generating a false positive match from grainy surveillance footage, leading to the wrongful arrest of an innocent person.
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
The incident was initiated by an incorrect match generated by the Clearview AI facial recognition system after its deployment, leading to a wrongful arrest.
Added categories peak at 2: Minor (Psychological Harm, Indirect)
A domestic law enforcement incident in which a false positive match from Clearview AI facial recognition software led to the wrongful arrest and four-year detention of an individual. While highlighting critical civil rights, privacy, and judicial transparency concerns regarding police use of proprietary AI, the incident has minor direct national security implications.
The AI system acts as an Assistant by analyzing surveillance images and suggesting 'possible matches' from a mugshot database to assist human investigators, who must then make the final decision to arrest or charge.
High Risk (Level 2): The report describes the use of facial recognition technology by law enforcement to identify criminal suspects, which is explicitly classified as a High-Risk AI application under the EU AI Act.
The AI system was used to analyze surveillance footage and match faces against a mugshot database to identify a suspect.
The report provides a highly detailed, legally documented account of [name withheld]'s arrest, the appellate court proceedings, and the specific role of facial recognition technology (Clearview AI) in his identification. The facts regarding his detention, the lack of discovery, and the systemic issues are clear and uncontradicted.
None. The causal link between the AI's false match and the subsequent wrongful arrest is clear and undisputed by the parties involved.
Technology
Flawed FRT Algorithms
Facial recognition software produced a false match from grainy footage.
Proprietary Software Secrecy
Developers require NDAs, preventing algorithmic transparency.
Data Inputs
Low Quality Surveillance Footage
Grainy images from store cameras increased the likelihood of a false match.
Historical Mugshot Databases
System matched suspect to old mugshots for unrelated past offenses.
Human Factors
Automation Bias
Police accepted a 'possible match' as definitive proof to arrest.
Confirmation Bias
Witnesses confirmed the suspect after being shown a biased photo array.
Process and Methods
Unregulated Out-of-State Searches
NJ police bypassed local protocols by using NYPD's database.
Withholding Technical Discovery
Prosecutors failed to provide algorithm details, violating due process.
Management
No Tracking of FRT Usage
The Attorney General's office did not monitor how many agencies used FRT.
Lack of Accountability Framework
Management failed to establish controls and safeguards before adopting FRT.
Regulatory Environment
Lack of FRT Governance
No clear statewide or federal laws regulate law enforcement use of FRT.
Interstate Cooperation Gaps
Cross-jurisdictional use allowed agencies to evade local discovery rulings.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
