LinkedIn Engineering

LinkedIn provides an update on testing how they measure bias in AI systems in the U.S. They discuss the challenges of measuring algorithmic bias and the limited availability of race and ethnicity information. LinkedIn outlines their approach, which focuses on effectiveness, privacy, and member control. They have developed a privacy-preserving probabilistic race/ethnicity estimation system to achieve their goals.

Responsible AI update: Testing how we measure bias in the U.S.

Challenges and Opportunities of Measuring Algorithmic Bias