Interpretive variability designates, in Pagup’s vocabulary, the observable dispersion of how AI systems describe, categorize, cite or recommend an organization across systems, queries, formulations, retrieved sources, moments or test contexts.
Why the term matters
An organization may receive a favourable answer in one context and a weaker answer in another. The issue is therefore not only whether it is visible, but whether its interpretation remains stable when the answer environment changes.
Interpretive variability helps identify where a brand remains readable and where it becomes fragile: displaced category, wrong competitors, missing proof, outdated offer, unstable recommendation or confusion with another entity.
Difference from AI perception drift
Variability first observes dispersion. AI perception drift then qualifies a gap against a baseline or canonical source.
In other words: every drift implies variation, but not every variation is drift.
When stability becomes suspicious
Interpretive variability does not only cover answers that change. It also forces analysis of answers that do not change. Apparent stability may come from delivery-layer fixation or stochastic fixation, when the answer delivered by an application no longer necessarily reflects a fresh generation.
How Pagup uses it
Pagup uses it as a diagnostic dimension in the LLM perception drift audit, AI representation stabilization and AI perception monitoring.
For the primary doctrinal definition, see gautierdorval.com. For normative qualification, see interpretive-governance.org.