AI Literacy

Neural Networks

Lots of simple connected parts that pass signals to each other, loosely inspired by the brain.

What's really going on

A neural network is made of many simple parts, called neurons, connected together. Each neuron follows one simple rule: if it gets enough signal from the neurons feeding into it, it fires and passes a signal on. No single neuron is smart, but a big network of them can recognize faces, understand speech, or play games. Modern AI systems like chatbots and image generators are built from huge neural networks with billions of connections.

For educators

Background

Artificial neural networks are loosely modeled on biological neurons. Each artificial neuron adds up its weighted inputs and fires (or produces a stronger output) when the total passes a threshold. The networks are arranged in layers: input, hidden, and output. Instead of being wired by hand, most networks learn by adjusting their connection weights to reduce errors on training data. Unplugged networks with fixed rules, where students pass tokens between neurons, make the layers and thresholds visible before students meet the idea of learning.

Misconceptions

  • A neural network is a digital brain that thinks like a person. It is loosely inspired by the brain but works very differently, and it has no understanding of what it’s doing.
  • Someone programs the rules for each neuron. In most real networks, people design the structure and the network learns its weights from data.
  • If you look inside a neural network you can see why it made a decision. Large networks are hard to interpret, and explaining their decisions is an active research area.