Previous layer Output 1 Next layer Weight 1 Output 1 x Weight 1 + Output 1 x Weight 1 Output 2 Output 3 Weight 2 + Output 1 x Weight 1 + Bias = Current layer’s output Weight 3 Bias Figure 5. Operation of an individual neuron Let’s take artificial neural networks as an example. A neural network is a large computational structure in which “neurons” are connected across multiple layers. Typically, every neuron in one layer is connected to every neuron in the next. This means each neuron receives as input the sum of the outputs from all neurons in the previous layer. Each of these connections has its own weight. A neuron is a function that takes one or more numerical inputs (the sum of outputs from the previous layer) and produces its own output; each neuron also has a bias value. After the bias is added, the resulting value determines what information is passed on to the next layer. In generative AI, the term “bias” can also refer to the phenomenon in which a model disproportionately reflects or excludes certain groups or perspectives due to its training data or design process. Examples Generative AI Guide for Civil Society