Love it. Once again, I puzzle looking for how this might position organizations to use AI to help reduce systemic disparities - transparently and explainably. Data bias is one sub-part of systemic oppression, and removal of bias is not an answer for those systemically excluded from classification.

At the same time, this is super high-quality conceptually. One can read into it answers to how organizations might take on their part of the ethical responsibility to reduce disparities over time in contexts they and their AI participate in, without discriminating.