Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
The United States Customs and Border Protection Technology Companies are asking a real-time facial recognition to send a pitch for a vehicle with any person in the rear seat, and according to a federal register last week, traveling documents in their documents.
Request for Information
“The entries that are biometically confirmed in the United States have been added to the crossing record of travelers,” the document states.
A company under the Department of Homeland Security, CBP says its facial recognition equipment is “currently working in the atmosphere of air, sea and land pedestrians.” The goal of the agency is to bring it to the “land car environment”. Accordingly A page on the CBP website Updated last week, the agency is currently “testing” how to do it. RIF says these tests prove that this facial recognition equipment is “advanced”, but it is not always able to get photos of every vehicle passenger, especially if they are in the second or third row.
“Human behavior, multiple passenger vehicles rows and environmental barriers present unique challenges for the car environment,” the document states. CBP says it wants to provide a private seller to a tool that “passenger images will increase” and “can capture 100% of the vehicles passengers.”
Dave Mass, Director of the Electronic Frontier Foundation Investigation, Has received a document The company published the results of the 12 -day examination by the CBP’s public record request from late 2021 to early 2022. Reported by the first intercept.
Mash said that what was standing to him was the rate of error. Unjalduas border crossing cameras on the Mexico border with Texas McLlen have just 76 percent of the car’s image and only 5 percent of these people have met “validation requirements” to match their identity documents.
The current repetition of the system matches a person’s photos in their travel document, known as one to one-mouth recognition. The initial risk here, MAS says the system is failing to recognize someone who matches their own documents. It is different from one to many facial recognition, which can use to identify a suspect based on police surveillance photo, where the initial risk is that someone is getting false positive matches and is being falsely identified as suspect.
MAAS says it is unclear whether the CBP error rates must be done to the camera or mill system. “We do not know that racial discrimination, gender discrimination, etc. Bring this system,” he said.
As report by Intercept in 2024, Department of Science and Technology of DHS Issued a request The same is similar to what CBP posted last week for information in August. However, the DHS document seems to be currently unavailable.