What is AI Governance?

A record of where data came from, how it was transformed, and where it is used.

How clearly an AI system's logic, data sources, and limitations are documented and understandable.

Human-added labels or metadata used to train or evaluate machine learning models.

A simple reference model used to compare whether more complex approaches actually improve results.

A standardized test or dataset used to measure and compare model performance.

How well a model's confidence scores match actual correctness probabilities.