Removing or masking the details in a dataset that could identify a person, so the data can be used without exposing who it describes.
Data Anonymization is a core preprocessing step for AI. Before data is vectorized or used for fine-tuning, automated pipelines need to remove PII, financial data, and classified IP so that the resulting model does not memorize sensitive information and leak it during inference.
See also Fine-Tuning, Air-Gapped Environment, Role-Based Access Control and Zero Trust Architecture.