In an era where technology has evolved to heights it has never reached before, facial recognition systems are one of the technology’s newest and latest achievements. While beneficial and helpful, it was found that this system also carried racial biases. Trained on disproportionate data sets, facial recognition systems are found to be terrible at identifying black faces – errors can come potentially incriminate the wrong person because of this.
Leonardo Selviaggo has the option besides properly proportioning the training data sets- elusion:
Selvaggio thought up the project, which he calls URME Surveillance
Selvaggio sees two routes to elude facial-recognition programs. The first is to disappear: go offline and off the grid. Selvaggio prefers the second option, which is to flood the system with weird, incongruous data. Wear someone else’s likeness or lend out your own. (Before donning a prosthetic mask, check to see whether your city or state has anti-mask laws, which may make wearing one illegal.)
Even without a mask, though, you can confuse some facial-recognition programs by obscuring parts of your face with makeup, costuming, hairdos and infrared light. Artificial-intelligence programs look for elliptical, symmetrical faces, so obscure an eye, cover the bridge of your nose, wear something that makes your head look unheadlike. “They have all of our information,” Selvaggio says. “So then let’s make more information that isn’t even true, and then let’s make more information on top of that.”
image credit: Nevit Dilmen via wikimedia commons