Managing bias in AI.

Roselli, D., Matthews, J., & Talagala, N. (2019). Managing bias in AI. WWW '19: Companion Proceedings of The 2019 World Wide Web Conference, 539-544. https://doi.org/10.1145/3308560.3317590

This paper addresses the challenges of managing bias in AI algorithms, particularly in light of the inherent risks for companies deploying these systems. The authors outline three general types of bias: those tied to translating business intent into AI, biases stemming from training data distribution, and biases within individual input samples. Despite the difficulty of eliminating bias entirely due to AI’s reliance on historical data, the paper offers best practices to mitigate its effects. The proposed processes focus on improving algorithmic outcomes through careful management of these biases.