I've been thinking a lot about ethics in machine learning and I've come to the conclusion that ML networks share their creator's biases. We should clearly mark that when a neunet messes up, it's the fault of the company that made it. When YouTube demonitizes LGBT videos for sexual content, that should be treated as an intentional decision by Google
@minty_cream unfortunately training is not scrutinized to the degree that the neural network is developed. We should be giving more emphasis to analyzing and validating training data sets since it's arguably the most important part of ML