Hello hello,
ICYMI – There’s some buzz in the tech world regarding the recent departure of an AI ethicist at Google. Timnit Gebru is a prominent black female scientist working in AI and best known for a paper she published in 2018 that found higher error rates in facial analysis technology for women with darker skin tones. She was co-authoring another paper which posited that tech companies could do more to ensure AI systems don’t perpetuate or exacerbate historical gender bias. Google wanted her to remove her name from the publication. Here’s an article about the unfolding event: https://www.theguardian.com/technology/2020/dec/04/timnit-gebru-google-ai-fired-diversity-ethics I’m sure there are many details left out from both sides in the story so I’ll hold off passing a judgement for now.
However, it shines a spotlight on the overarching topic of bias in technology. There’s a great book by Safiya U. Noble, Ph.D. called Algorithms of Oppression which outlines how search engines reinforce racism. Disclaimer: this book is dense! Is the search engine part of the problem or does the search engine merely reflect society’s beliefs? The book explores this distinction.
It’s one thing to combat systemic racism in the flesh, but it’s a whole different challenge to look at racial and gender biases in the technologies that we’ve incorporated into our everyday lives. The importance of this field of study will only become amplified as we continue to rely more heavily on algorithms across industries such as medicine, law, and manufacturing. It would be naïve to think that the field of AI and other technological advances would be free from these biases because they’re the creation of flawed humans working within a flawed system. They learn by evaluating human behavior and decisions, and we know human behavior and decisions are biased!! Acknowledging this fact is a critical step to put additional controls in place to combat these unwelcome biases in technology.