Bias refers to a systematic error or unfair trend in the data or model, leading to distorted outcomes. In computing, bias can arise from data collection or algorithm design, favoring certain outcomes over others. It's crucial to identify and correct for bias to ensure fair and efficient systems, similar to how overfitting must be mitigated in a model to enhance its generalization to new data.