21.1 Responsible AI

Responsible AI is the practice of designing, building, and operating AI systems in a way that is fair, transparent, accountable, and aligned with human values. It is not a single feature you add at the end; it is a set of principles applied across the whole lifecycle — from data collection and model training to deployment and monitoring.

For example, if an Ethiopian bank uses AI to approve loans, responsible practice means the bank can explain why an applicant from Bahir Dar was rejected, checks that rural applicants are not unfairly disadvantaged, and keeps a human reviewer who can override the model when needed.