Every introduction to neural networks says the same thing: they are inspired by the human brain. That framing is useful for about thirty seconds before it starts causing confusion.
Here is a more grounded way to think about it. A neural network is a series of math operations arranged in layers. Each layer takes a number as input, transforms it slightly, and passes the result forward. By the time a number has traveled through dozens of layers, the original input has been processed into something meaningful, like a classification or a prediction.
The neurons in this context are not biological cells. They are just functions that multiply an incoming number by a weight and add a bias. The word neuron is a metaphor that stuck, not a technical description.
What makes neural networks genuinely powerful is their depth. A shallow network might recognize edges in an image. Add more layers and it starts recognizing shapes. Add more still and it can identify objects. Each layer builds on what the previous one found, which is why this approach is called deep learning when the network has many layers.
The practical implication for a beginner: when someone says a neural network learned to do something, what they mean is that the weights inside those math functions were adjusted through training until the outputs matched the expected answers closely enough to be useful. No magic, just iteration.