Prediction consistency refers to the reliability and stability of the predictions made by a model or algorithm when subjected to small perturbations in input data or model parameters. Ensuring high prediction consistency is essential for building trustworthy models, especially in dynamic or high-stakes environments.