The principle
Adopters need a stable Standard they can reference in policy and training. Researchers and developers need room to improve instruments and programs as evidence grows. The policy lets the Standard change slowly and predictably while instruments and programs evolve at their own pace.
Versions
Every ALiF artifact uses a MAJOR.MINOR version number. For the Standard:
- A major version changes the structure (components, levels, role adaptation principles, or the number, names or fundamental definitions of sub-competencies) or adds, removes or tightens a conformance requirement. Major versions come with a consultation, a change log, a migration guide and a transition period.
- A minor version clarifies or refines without adding obligations, and is backward compatible.
A change counts as minor only if it passes four tests: no change to structure; no change to what assessment items measure, confirmed by expert re-rating where a definition changes; no new or removed conformance obligation; and only an editorial correction, a clarified boundary between sub-competencies, or an elaboration of something already present, recorded in a trace table. If any test fails, or the answer is uncertain, the change is major.
Major reviews
A major review normally happens every three to five years. An extraordinary review can be opened earlier on defined grounds, such as validation findings or a shift in the AI landscape. Each review follows set steps:
- a written, published record of why the review was opened;
- evidence synthesis;
- public consultation of at least 60 days, including the people affected by conformance requirements such as trainees and staff;
- a revised draft open for 30 days of comment;
- release with a change log that records how each comment was handled, and a migration guide;
- an 18-month transition period, with a migration plan due within 6 months.
Minor updates
Minor updates are reviewed each January and posted for 30 days of open comment before release.
Instruments and programs
Instruments and programs keep their own version numbers and state which version of ALiF they are aligned with. When a sub-competency definition changes, the self-assessment items are re-rated by independent experts before release. Items that fail are not scored under the new version, and definitions are not rewritten to fit items.
Version 2.0 of this policy
Version 2.0 was adopted at an early review on 1 October 2026 so that the review of the Standard for ALiF 2.0 follows these rules. The next scheduled review is January 2028.