Risk assessment identifies hazards, likelihood, impact, affected groups, controls, residual risk and accountable owners. Risk should influence autonomy and human oversight.
Responsible AI designs for human agency, fairness, transparency, privacy, safety and accountability across the lifecycle rather than applying an ethics checklist at the end.
Human oversight design specifies what people can see, decide, contest, override and repair. A nominal approval button is not meaningful oversight if the person lacks time or information.
