I wish there were a good solution to this issue, a very real and growing problem.

I believe ByteDance’s given justification of using the beauty algorithm to surface content is to prevent bullying (if we hide unattractive people, no one will make fun of them?), which is horrifying to Western sensibilities but makes a certain amount of sense if you divorce compassion for the overweight or non-traditionally beautiful…

But there are so many different things going on on top of this as well. The beauty algorithm is not only being used to identify beauty, but to identify individuals. When tied to the data that an app on a mobile phone is privy to, and the youthful decision making of the average user, that should scare us.

And, while it’s a more minor issue, I would argue that the beauty algorithm is tailored to traditional East Asian beauty standards, not Eurocentric ones. It’s a common misconception that Asian and Western beauty standards are the same, but having a waist like a snake (水蛇腰) or a small head, or v-line chins, even on men, have no context in the West.

White Westerners are often called “big noses” as a slightly derogatory but mostly silly term, but that’s something that the algorithm would ping white folk for (tall bridges in particular).

I don’t have a solution. But I do think this problem should serve for us as a warning in the development of future artificial intelligence tools: beauty is subjective, cultural, and has no merit (beyond aesthetic) in terms of content and value.

Thank you for discussing this topic: it needs to be addressed, loudly. And for counseling PoC not to worry about the beauty standards laid over their faces whether by callous artificial intelligences or catty classmates. I hope for a future when we can move past such judgment entirely — but until then, building up the confidence and emotional resilience of our children (and combating any notions of one race/mixed race being more beautiful than another) is the next best thing.

Thanks again.