5.2 Clause
Traditional AI
The family covers methods built from explicit procedures rather than learned parameters: rule systems, search, planning, constraint solving and optimisation. The behaviour is written by people and can usually be inspected rule by rule, which is its strength and its limit.
What this settles
Traditional methods are often the most auditable part of a modern system and the least visible in an inventory. Describing them properly restores the question of who wrote the rule and who may change it.
Who it binds
The organisation, as a method choice. In a regulated setting the rules may also be business rules with their own legal meaning, but that duty comes from the regulation, not from this draft.
What to do in your practice
For each traditional component, record the rule source, the owner and the change route. A rule base is a policy in executable form, and it should be versioned like a policy.
Evidence an auditor or regulator asks for
| A rule inventory with source, owner and version for each rule set. | Shows the rules are controlled documents. |
|---|---|
| A change record for a rule that altered a decision. | Proves the change route works. |
| A test set covering the rule boundaries. | Gives evidence for the behaviour claimed. |
Who is allowed to change a rule in this system, and is that change reviewed like a policy change?
Common gaps this page catches
- Rule sets live in code with no owner and no version history.
- A rule change alters a customer outcome and is never reviewed as a policy change.
- The inventory records "AI" and hides that the decision is fully rule-based.
What to prepare before publication
List the rule sets and give each an owner, a version and a change route. Start with the rules that touch eligibility, price, safety or a legal right.
How to test it here
Take one live system rather than a template, find the field or sentence where Traditional AI 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 family" and stops where the public record stops.
Common questions
- What does Traditional AI mean in ISO/IEC DIS 42102?
- The family covers methods built from explicit procedures rather than learned parameters: rule systems, search, planning, constraint solving and optimisation. The behaviour is written by people and can usually be inspected rule by rule, which is its strength and its limit.
- What evidence should we keep for 5.2?
- A rule inventory with source, owner and version for each rule set. A change record for a rule that altered a decision.
- Is Traditional AI 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