Item health

Which Fluentry test items are AI-co-authored, by test type. We publish the proportion of AI-assisted items per item type, never per item: individual item identity is intellectual property we do not surface. The figure refreshes weekly from our internal item-health view.

Last refreshed: 19 Jul 2026, 02:00 UTC. Refreshes every Sunday at 02:00 UTC.

AI co-authoring

Where AI contributed material, by item type. Pair items always carry an AI-authored side by construction, so their share is 100%.

Test	Share
AI Ethics	0%
Can AI Fool You?	18%
AI Knowledge	0%

For Knowledge and Ethics items, AI's role (suggesting wrong-answer options or phrasing variants) is recorded in our content repository rather than as a public per-item flag, so it is not yet counted here. AI never chooses a correct answer or sets a graded value.

What this shows

This proportion is drawn from the same item-health data we use internally to decide which items to keep, retire or replace. We publish the AI-co-authoring share so the claim on our methodology page is inspectable rather than asserted. The fuller aggregated item-health figures (bank size, accuracy, retirements) are available under NDA for procurement review.