Comment by mulmen

The danger is in thinking ML can magically see past human biases and/or is not subject to it’s own types of bias.