Researchers argue that, in some situations where machine-learning models are used to allocate scarce resources or opportunities, randomizing decisions in a structured way may lead to fairer outcomes.
Geigermeter News Magazine