Skip to content

Reference & glossary

TermMeaning
AI system impact assessmentStructured evaluation of how an AI system and foreseeable applications may affect individuals, groups or society
Affected partyPerson, group or community experiencing direct or indirect effects
Interested partyPerson or organisation that can affect, be affected by, or perceive itself affected by the assessment
Actual impactImpact already observed
Reasonably foreseeable impactPlausible impact supported by context and evidence, not remote speculation
MeasureTechnical or management action addressing harm or benefit
Residual impactImpact remaining after measures
MaturityHow established and managed an assessment practice is
AssuranceHow strongly the assessment result has been verified
Approved N/AHuman-approved decision that an item is irrelevant to the stated system context

It is guidance for AI system impact assessment. Gamut does not issue certification.

The module separates the process and documentation guidance into unique atomic checks so assessors can identify exact gaps rather than score broad paragraphs.

No. A mature process may identify a severe residual impact and decide not to proceed.

Does stakeholder engagement determine the decision?

Section titled “Does stakeholder engagement determine the decision?”

It informs accountable judgement. The assessment should record differing views and explain the decision.

No. Missing information is uncertainty or a gap.

Does Privacy Mode remove affected-party data from AI prompts?

Section titled “Does Privacy Mode remove affected-party data from AI prompts?”

It excludes direct identifiers and free-text content, sending only bounded structural state.

Gamut provides a system-specific assessment aligned to ISO/IEC 42005:2025 guidance. The result is an accountable impact-assessment record, not certification or a guarantee that all impacts have been eliminated.