This is a working note.
Ethical AI should be evaluated by how it affects human judgment, privacy, accountability, learning, and institutional trust.
The useful question is not whether an AI system sounds impressive. It is whether the system helps people make better decisions while preserving the conditions that make those decisions accountable.