Quiz 1 - Question 12

I have a remark on this question (image attached).

The question is phrased “Which of the following are likely?”. I find it highly unlikely that a model is ever able to remove ALL avoidable bias. I agree with the statement that it would be difficult to identify this bias, but surely it is still likely that a bias is present? Albeit an extremely tiny one? Please tell me if I am crazy here :slightly_smiling_face:

Please read the error response again: they are not claiming that there is no possible bias, but the question is how you define or identify bias once you have done better than human error. I admit that maybe we should be applying to law school here, instead of doing machine learning, but that’s the game we are playing here …