5 Clause
AI system methods dimension
The clause presents the method side of the matrix. It separates four families, traditional AI, symbolic AI, machine learning and hybrid AI, and gives each its own descriptor set. The families are not a maturity ladder: a rule engine is not a worse neural network, it is a different method.
What this settles
The method axis is the part of the framework that stops AI being treated as a synonym for machine learning. A system built from search, constraints or logic still needs describing, and often needs more explanation than a model, not less.
Who it binds
The organisation, as a method choice. No legal duty arises from the clause. Its effect is that a team cannot hide a mixed or rule-based system inside a generic machine learning label.
What to do in your practice
Start the method register with the four families and add a fifth column for "mixed". For each system, record which families are present and which descriptor in the family applies. A family with no entry is a finding, not a blank.
Evidence an auditor or regulator asks for
| A method register with one row per system and one column per family. | Shows the method position is recorded, not assumed. |
|---|---|
| A descriptor note for each family a system actually uses. | Gives a reviewer the detail behind the label. |
| A list of systems with no method recorded. | Makes the gap visible and assignable. |
How many systems in our estate use no machine learning at all, and would our inventory show them?
Common gaps this page catches
- The inventory equates AI with machine learning, so rule-based and symbolic systems disappear.
- A system is labelled hybrid without saying which methods are combined.
- The method family is recorded once and never revisited after a rebuild.
What to prepare before publication
Add the four family columns to the inventory and fill them for every live system. The exercise usually changes the count of AI systems a firm believes it has.
How to test it here
Take one live system rather than a template, find the field or sentence where AI system methods dimension 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 "methods dimension" and stops where the public record stops.
Common questions
- What does AI system methods dimension mean in ISO/IEC DIS 42102?
- The clause presents the method side of the matrix. It separates four families, traditional AI, symbolic AI, machine learning and hybrid AI, and gives each its own descriptor set. The families are not a maturity ladder: a rule engine is not a worse neural network, it is a different method.
- What evidence should we keep for 5?
- A method register with one row per system and one column per family. A descriptor note for each family a system actually uses.
- Is AI system methods dimension 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