Raw LLM Responses

The machine learning models aren't trained on non-whites, so they end up being inaccurate to the point of uselessness: https://www.wired.com/story/best-algorithms-struggle-recognize-black-faces-equally/ It's just like how human beings have difficulty recognizing and distinguishing other-race faces. Unless you spend a lot of time around ethnic Chinese or Japanese, for example (assuming you're an average Redditor: white American, Canadian, or British), you probably aren't as good at telling them apart as you are at people if your own race who you interact with more frequently. It's not maliciously motivated in either case, but the outcome in the case of has AI is racially weighted and therefore untenable.