Good, bad, or ugly, it is a historical document written at a time when it "seemed like a good idea." AI is going to be AI and how you train it is what it does.<p>There should be nothing surprising about an AI program detecting the ethnically abusive language in the DoI and flagging it.<p>Did they fix it? Yes. Does this make the whole DoI worthless? No, and we should be fit to learn from mistakes of the past and treasure what is good.<p>*edit for formatting