Artificial neural networks are (most likely) not like the human brain
When teaching or learning neural networks, many fall into the trap of thinking that the perceptron is a mathematical model of a neuron. This is wrong. We do not really understand how the brain works (despite some advancements in the field), and therefore we have no idea if a neuron works like a perceptron — most likely, it does not.
What seems to be true is that the perceptron was inspired by how an individual neuron works.
Wrong: a perceptron is a mathematical model of a neuron.
Right: the perceptron was inspired by the neuron.
This misunderstanding generalises to neural networks: when people assume that a perceptron works like a neuron, they also assume that a neural network works like the brain. Again, this is wrong!
Wrong: an ANN works like the human brain.
Right: an ANN does not work at all like the human brain (for all we know).
This misinformation spreads, especially among less technical people. This can cause undesirable consequences, like believing that we just need a better computer and we will replace the need for human work.
When you teach or tell about neural networks to someone, make sure you explain the (subtle) difference between being inspired by something and being a model of something. The perceptron is not a model of a neuron and ANNs are not a model of the brain (for all we know).
Action: when you explain ANNs/perceptrons, make sure to explain this subtle difference as well.