Definition
decision to use an AI output for a defined purpose after its relevant claims, limits and permissions have been assessed by a responsible person
AidGPT working definition. Source basis below.
Notes to entry
Note 1 to entry: AidGPT's course asks learners to judge when they may rely on AI work. A coherent answer or the system's own justification does not establish that its claims are supported. Review the output, verify consequential claims against independent evidence and state remaining uncertainty.
Note 2 to entry: A sound output may still be unsuitable for the audience or outside the permitted use. Identify who can authorise the next action. Human review, verification and approval are separate steps and must not be recorded as though one completed the others.
Note 3 to entry: The amount of checking depends on the task and consequence. A decision not to use the output can be justified. Record the reason and the point at which the work should return to a person.
Note 4 to entry: NIST AI RMF 1.0 calls for documented system limits, intended use, expected benefits and costs, and human oversight. The wording of this entry is AidGPT’s course-specific synthesis, not a term defined by NIST.
Source basis
AidGPT working definition. These sources informed the concept and its scope. They did not issue or approve this wording.
- DCI AI Hub, AI Hub Risk Assessment Framework for Social Protection (Sector risk framework)
- Core Humanitarian Standard on Quality and Accountability, 2024 (Humanitarian sector standard)
- NIST AI Risk Management Framework 1.0, Core (National risk framework)