A UNICEF Giga participant asks how data foundations and application harnesses should be prioritised together.
Question raises complementarity; no comparative test.
Speaker
Unidentified speaker
Making AI Useful in Humanitarian ResponseWatch passage →
DEFINITION USED IN THIS COLLECTION
Assurance learning.
Continuing assurance, monitoring, feedback and improvement over a system's life.
What it includes
Continuing assurance, monitoring, feedback and improvement over a system's life.
Where the boundary lies
A one-time origin record does not establish ongoing assurance.
How speakers put it.
The definition above is used to organise this collection. The passages below show how the idea appears in discussion.
Evidence in practice.
All 31 coded passages →Analyst applies. A UNICEF Giga participant asks how data foundations and application harnesses should be prioritised together.
Qualifies the boundary. A one-time origin record does not establish ongoing assurance.
A packaging workflow extracts claims, maps them to a taxonomy and issues evidence requests for verification.
No error analysis or production outcome shown.
Speaker
Multi-stakeholder, Sovereign, Agentic Workflows / Data Models, Evidence, and VerificationWatch passage →
Differences and limits.
No event-wide agreement is inferred from coding. Compare the attributed examples and boundaries; co-occurrence is not evidence of agreement.
A one-time origin record does not establish ongoing assurance.
Related concepts.
evaluation →
Evaluation concerns judging performance or effects; assurance learning emphasises continuing monitoring and improvement.
Codes that occur in the same passages
Co-occurrence describes coding, not equivalence, causation or agreement.