Joseph Gatto's answer to What are some bad use cases of neural networks? - Quora

Also, neural networks are very difficult for humans to interpret. You may be in a situation where it is important that you know exactly what factors in your model lead you to your results. That can be extremely difficult with a neural network considering the massive amount of factors that go into a neural network's output. Simpler ML models can show more precisely what factors lead to your results.