AidGPT

How the definitions are written

One concept per entry, with the definition, explanatory notes and evidence kept distinct.

Some entries quote named standards or a NIST glossary. Others adapt a source or state AidGPT working wording. Each entry identifies which applies. The collection does not claim ISO conformity.

Method

Each terminology page describes one concept. Its definition names a broader kind of thing and the feature that distinguishes this concept. Numbered notes explain scope, use and practical cautions. Notes do not silently add a condition to the definition. A term used in a definition links to its own entry. The concept relations section shows broader and related concepts, while the source list shows the material used to frame AidGPT’s wording.

This approach draws on ISO 704:2022 for concept-oriented terminology work and ISO 10241-1:2011 for the structure of entries and notes. ISO 1087:2019 supplies the terminology-work vocabulary. The complete standards contain more detailed requirements than this website claims to implement.

Source status

A citation alone does not make a definition a standard. Each entry identifies whether its wording is quoted, adapted or written by AidGPT. The source list also names the kind of publication. These distinctions matter when a technical term is used in a course, policy or assessment.

Entry wording What the source establishes
Quoted definition The short wording comes from the named publication and clause or glossary entry. Notes are AidGPT additions.
Adapted definition AidGPT has shortened or recast a named source explanation. It is not the source's formal definition.
AidGPT working definition AidGPT has written the wording for its course or humanitarian taxonomy. Listed standards and frameworks inform the concept, but did not issue this definition.

The traditional AI label is informal. The course's precise comparison is with rule-based automation. A technical explanation from IBM informs those entries but is not an industry standard.

Concept map

Humanitarian AI joins an AI system to a task in humanitarian action. A humanitarian AI use is a particular application with a purpose and setting. Describe what the system does through its capabilities and why the work is undertaken through its function. Then examine who may be affected, the system’s permissions, the risk assessment, oversight and any approval.

Limits

The AI system, AI model and AI output entries quote short definitions from named ISO/IEC publications. The generative AI entry quotes a NIST glossary definition. The AI output source uses the shorter term “output”. Capability entries adapt explanations from the DCI/GIZ taxonomy for social protection and identify the section used. AidGPT’s humanitarian function labels are proposed adaptations, not recognised sector codes. Each entry labels its definition as quoted, adapted or AidGPT working wording. The linked sources do not endorse AidGPT’s notes or adaptations. A definition helps people describe a use. It does not establish that a system works, that its risks have been addressed or that anyone has authorised it.

Sources