Do you believe that facial recognition should be banned even if the error rates went away? Is there any other huge disadvantage that outweighs the advantages?

Hi! I’m Seanna, and I’m completing my senior year at Bard High School Early College Queens. I live in Queens and I love ice cream (salted caramel is amazing). In our Internet and Society class, we have been reading your work, and I chose your piece on facial recognition to focus on.

In your article, “A Case for Banning Facial Recognition,” you show the reasons a scientist, Timnit Gebru, believes facial recognition should be banned. Gebru talks about how based on research this technology can be misused and potentially target dark-skinned women as opposed to light-skinned men. This really interested me so I did further research. A study on different facial recognition software showed all of them had much higher rates for dark-skinned females than light-skinned males, some went as high as 34%. Gebru also mentions this idea of automation bias, where if a computer tells you something you tend to believe it, even if it goes against your instincts. I think this makes complete sense which is why it is dangerous to be so reliant on facial recognition even when there are high error rates.

Near the beginning of the article, you mentioned how “civil liberties groups say that facial recognition contributes to privacy erosion.” The main reason Gebru believes that the technology should be banned is because of the high error rates that feed into racial bias. The error rates outweigh any advantages that the software presents. Do you believe that facial recognition should be banned even if the error rates went away? Is there any other huge disadvantage that outweighs the advantages?