Which ML algorithms are racist? I could see how a researcher might train on a data set with built-in biases and assumptions, but I'm genuinely struggling to see how something like linear regression or SVM are "massively racist". Or maybe you are being sarcastic, I can't tell.
It’s the data, not the algorithm.
I'm using "algorithm" to mean the complete usable product including training not the contextless regression itself. Come on man.
Algorithm has a specific definition. That is not it.