Machines have this nice property that they don't make mistakes.
They only repeat the mistakes of their designers, but they never make new mistakes.
This is good. This lets us reason about them, debug them, and rely on them always behaving the same way.
When you use Machine Learning, you lose that property. You turn a predictable machine into an unpredictable monkey.
@Wolf480pl but unpredictables monkeys still have useful application, am I wrong?
We are a cute and loving international community Ｏ(≧▽≦)Ｏ !