Model Drift

Model drift occurs when an AI model’s performance drops over time because real-world data has changed. In voice platforms, this might mean new phrases or accents the system didn’t see during training. Regular updates and monitoring are needed to keep accuracy high.