Artificial neurons are the fundamental building blocks of artificial neural networks, inspired by the biological neurons in the human brain. They process input data through weighted connections and activation functions to produce an output, enabling complex computations and learning in machine learning models.