What is data lineage in AI data management?

• A. A method for anonymizing data.
• B. A dataset labeling process.
• C. The end-to-end trace of data from origin through transformations to outputs; supports trust and auditability.
• D. A model training technique.

Answer: (C)

Data lineage is the end-to-end record of where data comes from, how it’s transformed along the way, and where it ends up in outputs. In AI data management, this visibility lets you trace model results back to the original data and every processing step—cleaning, feature engineering, joins, and aggregations—so you can assess data quality, ensure regulatory compliance, and reproduce or debug models. If a prediction seems off, you can follow the trail from outputs back to source data and transformations to understand what drove it. This contrasts with options like anonymizing data, which is about privacy masking; labeling data, which is about tagging data; or model training techniques, which describe how a model learns—not how data flows and is transformed.