The article that struck me most was Nabil Hassein’s “Against Black Inclusion in Facial Recognition”. I’ve always heard and seen people of color passionately pushing for more diversity and representation in artificial intelligence to counteract algorithmic bias, like Joy Buolamwini and Timnit Gebru and thus I also began advocating for the same. But Hassein’s take on how we’ve seen “unbiased” facial recognition software continue the legacy of wrongful accusation and criminalization of Black people caused me to take a step back. Hassein explains: “…by refusing to don white masks, we may be able to gain some temporary advantages by partially obscuring ourselves from the eyes of the white supremacist state”. I’d never thought of algorithmic bias against Black people as a positive side effect, however I still wonder what remains to be done for those who have been wrongfully convicted by facial recognition. It seems far-fetched that we would be able to eradicate law enforcement’s use of AI, so I think it’s important for our focus to go to making the systems less biased to stop even more people from being framed.
In Kyle McDonald’s course notes on “Face as Interface”, I learned more about the principles and terminology of facial recognition technology. I also learned about an interesting piece of information where McDonald describes a phenomenon in Japan where customer service workers develop emotional trauma from something called “smile mask syndrome”. This results from workers having to force a positive attitude in a society that places a significant value on exceptional customer service. This connected to McDonald’s earlier claim that facial expressions actually generate emotions and affect the nervous system. Although I’ve heard people say that smiling actually makes you feel better, I’d never considered the flip side where you’re smiling and faking emotion so much that it throws your nervous system into distress, and it seems very plausible. This information does not appear to be widely know which makes me concerned that workers in certain industries are experiencing a degree of suffering that may not be treated and taken seriously.
Voice recordings incase you need to test it:
Joy Buolamwini discusses algorithmic bias in her TED Talk. The way facial recognition software generally works is that they are given a “training set” which tells the machine what is a face and what is not. This gives the machine the ability to detect faces. However, the training sets are not always inclusive of all people. Despite this, I especially liked how she phrased this issue by emphasizing that more inclusive training sets must and can be made.
