7.3 Clause
Characterization framework of AI system methods and capabilities
The subclause presents the matrix itself: methods against capabilities, with a descriptor set on each axis. The matrix is the artefact a team fills, and the draft says a system can hold several methods and switch between them while performing a task.
What this settles
The matrix is only as good as its evidence discipline. Filled cells should mean "claimed and supported", empty cells should say why, and the whole grid should be revisable when a method changes.
Who it binds
The organisation, as a method choice. The clause creates no duty, but a matrix used in procurement or assurance becomes part of what the organisation has asserted.
What to do in your practice
Give every claimed cell an evidence pointer and a review date. Mark empty cells as not applicable, not yet evidenced, or outside scope. Review the grid when a model, method or capability changes, and keep the previous version.
Evidence an auditor or regulator asks for
| A matrix with an evidence pointer and review date per claimed cell. | Makes every claim traceable. |
|---|---|
| An empty-cell key with three states. | Stops "not done" being read as "not applicable". |
| A dated previous version after a method change. | Shows the description follows the system. |
Which claimed cell would fail if we had to produce its evidence today?
Common gaps this page catches
- Cells are coloured by confidence rather than supported by evidence.
- An empty cell is read as "not applicable" by every recipient.
- The matrix is never versioned, so no one can tell what changed after a rebuild.
What to prepare before publication
Add the evidence pointer, the review date and the three-state empty key to the template. Then test the matrix on the system with the most change in the last year.
How to test it here
Take one live system rather than a template, find the field or sentence where Characterization framework of AI system methods and capabilities is used, and put this page beside it. Record the answer, the artefact and the date. A glossary entry on its own is an intention; a dated record beside a live artefact is evidence.
What this page does not claim
The page does not reproduce the draft, does not know the descriptor wording that the public preview withholds, and does not certify the system or the organisation. It states the unit at the claim level "framework clause" and stops where the public record stops.
Common questions
- What does Characterization framework of AI system methods and capabilities mean in ISO/IEC DIS 42102?
- The subclause presents the matrix itself: methods against capabilities, with a descriptor set on each axis. The matrix is the artefact a team fills, and the draft says a system can hold several methods and switch between them while performing a task.
- What evidence should we keep for 7.3?
- A matrix with an evidence pointer and review date per claimed cell. An empty-cell key with three states.
- Is Characterization framework of AI system methods and capabilities a requirement?
- No. The draft is a descriptive framework. It sets no management system requirement and no conformity test. It becomes binding only when a contract, a regulator or a management system adopts it.
Related units
Last updated: · Markdown version · Check your system