Terms related to trustworthiness · 3.5.8
predictability
a property of an AI system that lets stakeholders make dependable assumptions about its output
What this term settles
The term looks ordinary until two teams use it differently. The entry removes that room by naming the thing, its inputs and its limits in one sentence.
Who it binds
The clause does not instruct anyone: it settles what predictability means so that an instruction written elsewhere can be read the same way by both sides.
What to do in your practice
Take the firm's own definition of predictability, put it beside this one and mark the differences. Each difference is either a deliberate local choice or an inconsistency to fix.
Evidence an auditor or regulator asks for
- A controlled vocabulary entry for predictability with the citation 3.5.8, an owner and a review date.An auditor can check that predictability is controlled rather than assumed.
- An assurance argument that links the property to the test, the review or the incident record behind it. The row names predictability.Shows predictability in use in the artefact the firm relies on, not only in a glossary.
- A register row with the owner, the date and the evidence reference for the claim. The record carries the predictability entry and its date.The assurance case asserts the property without naming the test, review or incident record that supports it. The gap is tested against predictability.
Common gaps this page catches
- predictability is used in the assurance argument without a written definition, so two teams can mean different things.
- A supplier uses predictability with a different boundary and nobody has recorded the difference.
What to prepare before the amendment lands
Brief the owners who use predictability most often, then ask each to return one artefact that shows the entry in use. File the artefacts against the register row.
How to test it here
Take one live artefact rather than a template, find the sentence or field where predictability is used, and put this page beside it. Record the answer, the artefact and the date, because that record is the evidence and a glossary entry on its own is only an intention.
What this page does not claim
The page does not reproduce the standard, certify the firm or predict the wording the amendment will use; it states predictability at the claim level "definition" and stops where the public record stops.
Common questions
- What does predictability mean in ISO 22989?
- a property of an AI system that lets stakeholders make dependable assumptions about its output
- What evidence should we keep for 3.5.8?
- A controlled vocabulary entry for predictability with the citation 3.5.8, an owner and a review date. An assurance argument that links the property to the test, the review or the incident record behind it. The row names predictability.
- Is predictability a requirement?
- No: predictability is a definition or clause of a terminology standard, and it becomes binding only when a firm writes it into a policy, a contract or a procedure, or when another standard incorporates it.
Related units
Source. ISO/IEC 22989:2022, Clause 3.5.8, Terms and definitions. Open the ISO public entry. Retrieved 21 October 2026.
https://www.iso.org/obp/ui/en/#iso:std:iso-iec:22989:ed-1:v1:en
ISO catalogue record 88145, the amendment status, stage and dates. Catalogue record 74296, the base standard.
Claim level: definition. The definition is in the public preview. This page states it in our own words and does not reproduce the text.
Last updated: 2026-10-21 · Markdown version · Check your vocabulary