AI Fairness 360 is best described as what?

• A. An open-source toolkit to examine, report, and mitigate discrimination/bias in machine learning models
• B. A commercial data labeling service
• C. A private benchmark dataset
• D. A cloud hosting platform for AI workflows

Answer: (A)

AI Fairness 360 is an open-source toolkit designed to examine, report, and mitigate discrimination and bias in machine learning models. It provides a set of fairness metrics to diagnose bias, along with algorithms and workflows to mitigate that bias at different stages—data, models, or predictions. Because it’s open-source, you can run it locally, compare different fairness definitions, and reproduce results across projects, which makes fairness work more transparent and actionable.

This focus on measuring bias, offering concrete remediation methods, and being freely accessible distinguishes it from the other options. It isn’t a commercial data labeling service, a private benchmark dataset, or a cloud hosting platform for AI workflows, so those choices don’t describe what the tool actually is.