My project is made up of various personal photographs of myself, my friends and family. I modified these images to hide the identity of the people photographed, and later used different facial recognition programmes to see if the AI could still detect faces or tell anything accurate about the people photographed. Through these photos, I explore identity in the context of modern technology.
I was inspired by Trevor Paglen’s and Kate Crawford’s 2019 viral project ImageNet roulette. ImageNet is a highly trained AI that analyzed pictures uploaded to a website. The results revealed how our biases and judgement against other people can easily live on many years from now as we continue training and programming different AI. The difference seems to be that humans have some kind of moral code and can condemn such judgements, whereas computers just continue to do what they’re told and share the information they’ve been given.
Because of lockdown, we can’t really see each-others faces anymore. Social media and video calls have made it possible for us to talk face to face, and i’m not sure whether talking to someone from 2 metres away while wearing a mask can really compare to that. In a way, we imitate the way facial recognition technology tries to make deductions from heavily distorted images, by trying to read someone’s expressions from behind a mask.