In the space of three days this month, two local governments walked away from the same kind of system.
On 6 October, Henrico County, Virginia, ended its automated license plate reader contract early. The county had installed more than 120 of the cameras since November 2022 and spent about $1 million on them. County Manager John Vithoulkas put the problem plainly: “it seems like every day, there’s another issue with another locality and how Flock cameras are used.” Two days later, Reno said it would not renew its agreement with Flock Safety, after residents argued that cameras logging every passing vehicle into a searchable database violated their privacy and constitutional rights.
Neither was first. Reporting in January counted at least six jurisdictions, including Flagstaff, Arizona, Cambridge, Massachusetts and Evanston, Illinois, that had cancelled or terminated contracts. Some cited residents’ privacy concerns. Cambridge and Evanston cited cameras installed or reinstalled without the city’s knowledge.
Look at the reasons given in Reno, Henrico and the cities before them and a pattern shows up. The complaint is not that police find stolen cars or wanted people. The complaint is the method: record everyone who passes, keep it, and search it later.
Henrico’s system kept images for 21 days. Every driver who passed a camera in that window, suspected of nothing, sat in the searchable set. That is the design choice that keeps losing votes. Not the goal. The collection.
There is another way to build identity technology for public safety, and it begins at the opposite end. Instead of collecting everything and searching it later, you start from a short list of records a court or agency has already created, and you ask a narrow question about a single moment: does this person match one of them?
That is the model Safience is built on.
Records first. Every comparison runs against UMbRA, the law-enforcement-sourced identity database, which holds 58.4M+ records and draws on nothing else. No social media, no news photos, no commercial data.
Nothing kept on anyone else. A facility captures a single still image. If it does not match a relevant record, it is deleted at once and no biometric template is stored. There is no history of who walked past.
No video, no plates. Safience never receives, records or stores video. It does not read license plates or follow vehicles.
A person before the alert. A trained analyst reviews every potential match before anyone is told. A resemblance is not treated as an identity.
Reno staff plan to bring a contract with a different vendor to the council on 21 October. Whatever that council decides, and whatever any other council decides, the questions that tend to settle public comment are the same.
A vendor that answers the first two with “everyone” and “weeks” is asking a council to defend mass collection at the microphone. Recent votes show how that tends to end.
Public safety tools are losing votes because they collect too much, not because residents are against public safety. A system that starts from a verified record, keeps nothing on anyone else and puts a trained person between the match and the action gives a council something it can stand behind.
For how a single record-bound match works in practice, see Identity intelligence starts with a record, not a face. For how biometric rules differ by state, see our state privacy policy pages. Agencies and councils with questions can reach us through contact.
Sources
Featured photo by Michael D Beckwith on Unsplash.
A face search can do its job perfectly and still put the wrong person in...
Read articleA stop or an arrest that starts from an identity match fails in one of...
Read articleBy a Biometrica staffer Drug related offenses accounted for almost one-third of all cases reported...
Read article
