diverse and representative group
Mitigating AI Bias, with …Bias
This article is part of my Data Trust series of talks and writing. The purpose of these articles are to break down complex but important socio-technical topics in a manner that is accessible to both practitioners and non-practitioners. Most tools we use today leverage AI/ML from the moment we wake up and while we sleep. Humans build Machine Learning, and humans are inherently biased. Since humans aren't perfect, we encode our biases into the data we use to train AI.