Executive Order 13769, Protecting the Nation From Foreign Terrorist Entry Into the United States issued in the first days of the Trump presidency, called for the Secretary of Department of Homeland Security (DHS) to expedite the completion and implementation of a biometric entry-exit tracking system for all travelers to the United States. To date this system has been implemented at 15 major airports and points of entry with a goal of 21 by 2021. This expanded deployment by Customs and Border Patrol has heralded an explosion of the technology’s usage by other agencies particularly law enforcement which are taking advantage of faster computers, better algorithms and more images to compare against in an ever-growing database of images.
The adoption of these tools has raised questions about their accuracy and citizen privacy. Municipalities such as San Francisco and the Boston suburbs of Brookline and Somerville have banned their use by agencies under their control.
Biometric facial recognition software reads the geometry of your face; the distance between your eyes, distinguishing characteristics like moles or birthmarks, the shape of your nose, some systems will define up to 68 different characteristics and convert this into a digital signature. This signature can be compared against existing databases that have already been processed; FBI mug shots, passport photos, terrorism watch lists and compared for matches. With the explosion in procesing power this can be accomplished while standing in an arrivals or departures terminal in an airport.
The technology has many uses beyond screening people entering or leaving the country and preventing undesirables from traveling. Police departments can identify suspects by photos from surveillance cameras or bystander photos. Merchants can identify known shoplifters before they steal anything. Sexual predators can be removed from the mall before they even enter. Conceivably this could replace or supplement drivers licenses, photo IDs or even library cards.
Here’s where things get interesting. In our personal frenzy to partake in social media and digitally document every aspect of our lives, we have voluntarily uploaded billions of photos of ourselves, our families and friends with accompanying personal information; names, locations, dates and associations. Whereas organization like DHS are committed only to using federally authorized databases like passport and FBI photos, other organizations are not similarly constrained. The New York Times on January 19, 2020 reported on a startup called Clearview AI that has created and marketed an app and database that not only searches public photo records but Facebook, YouTube, Venmo and millions of other websites. Facial recognition works best when there are more images to compare against. A single image can produce a false positive. Verifying against multiple images make false positives less likely.
In a previous USResist News post, congressional legislative efforts to curtail abuses of this technology were explored. These efforts have been limited to law enforcement agencies and have, regardless, fallen short in the Senate. Without direction on the national level it seems clear that other entrepreneurs will seize the opportunity to monetize and weaponize this technology assembling ever larger sources of video and photographic data to further the accuracy of searches.
Banning the technology will not stop its use but regulating how companies can build their databases might provide an avenue to contain abuse.
- The Markula Center for Applied Ethics at Santa Clara University has prepared a report Facial Recognition Technology An Ethics Case Study that explores the ethical considerations for biometric facial recognition
- The American Civil Liberties Union is leading the fight to curb abuses of Biometic screening.
- 30 diverse organizations that believes that facial recognition poses a threat to human society and basic liberty that far outweighs any potential benefits.have banded together in com to present a compelling argument for regulation
Photo by Andy Lee