AI Ethics

How can we tell whether an AI system is being used responsibly?

Principles such as fairness and transparency become useful when connected to a specific use, affected people and a way to detect failure. We can ask who defines acceptable performance, who bears the risk and who can stop or correct the system. Ethical labels cannot replace evidence. Different uses require different safeguards, and uncertainty should be visible to those relying on the result.

A question to sit with (optional): Who can contest a harmful result, and does that challenge reach someone able to act?