North Florida Sheriff’s Office Identifies an Uncooperative Felon
BENEFITS
87% faster identification time than other methods that didn’t work
Independently confirmed suspect identification in 2 minutes
Helped increase
officer safety
PROBLEM
INDIVIDUAL REFUSING TO IDENTIFY THEMSELVES
In December 2019, a Florida Sheriff’s Office patrol officer pulled over a driver and her passenger. The passenger was uncooperative in providing his ID, provided two different names and dates of birth, and claimed not to know his Social Security number. Tattoos on the passenger’s neck indicated a possible gang affiliation. From his office, a Crime Analyst Supervisor started to assist the patrol officer in uncovering his identity.
The Crime Analyst Supervisor first tried to search a text-based database of arrestee scars, marks, and tattoos. After 15 minutes, he could not find a match of the passenger’s tattoos in the database.
Next, the analyst searched a photo of the passenger taken with the patrol officer’s smartphone against Florida’s state-wide facial recognition database. After another 15 minutes, and even after cropping and enhancing the photo, he still could not find a match for the passenger.
“With the other facial recognition programs, sometimes I have to crop the picture or I have to lighten it up…. With Clearview AI, I didn’t have to do anything to it.”
SOLUTION
IDENTIFYING A SUSPECT AFTER OTHER METHODS HAVE FAILED
Facial recognition solutions based upon artificial intelligence can help law enforcement agencies identify individuals when other methods have failed. Agencies can use the facial recognition results as investigative leads, and when supported by other evidence can accurately and rapidly identify suspects, persons of interest, and victims of crimes.
Running out of time and options, the analyst submitted the uncropped, unenhanced photo of the passenger to their new facial recognition tool, Clearview AI.
Clearview AI is a post-event research tool for law enforcement, using over 50 billion photos derived from publicly available web sources, including social media posts, personal and professional websites, news articles, online mugshots and other criminal databases, public record sites, and thousands of other open source records.
“Within three seconds, I got four positive hits. Three of them were different versions of the same mugshot. The tattoos were clearly visible, the same hairline, the same eye set, the same eyebrows.”
THE RESULTS
CLEARVIEW AI HELPS TO IDENTIFY A CONVICTED FELON
Within three seconds, Clearview AI found four results, all pointing to the same person. The most helpful result was a previous arrest record found online including his name, date of birth, and arrest history. The analyst followed this investigative lead by utilizing his facial recognition training, and cross-checking the information with the Florida Department of Corrections. Within under five minutes, he had helped determine the former inmate was the passenger in the car.
After the analyst identified and confirmed the suspect’s identity through further investigative methods, he then informed the patrol officer, “I think I got him,” and gave the patrol officer the passenger’s name. The patrol officer addressed the passenger with his real name, which the passenger confirmed, ending the stalemate.
Further investigation determined that the passenger had served time on multiple charges, including fleeing and eluding a law enforcement officer and was wanted on multiple warrants. The patrol officer arrested the passenger and brought him to jail.
Without Clearview AI, more time would have passed with the patrol officer waiting for a mobile fingerprint reader to help identify the uncooperative passenger. Since he had already served time for fleeing and eluding an officer, there is no guarantee that this classified violent felon would still be there by the time the mobile fingerprint reader arrived.
