20 AI Hallucination Check Templates

Claim ledger: Claim → [ ]; risk → [low/medium/high]; source → [ ]; status → [verified/corrected/qualified/removed].

Citation check: Source exists? [ ]; exact passage supports claim? [ ]; date/version current? [ ]; population/jurisdiction matches? [ ].

Quotation check: Exact wording → [ ]; speaker/author → [ ]; primary source → [ ]; location → [ ]; punctuation/translation caveat → [ ].

Final AI verification pass: names [ ]; dates [ ]; numbers [ ]; quotations [ ]; citations [ ]; links [ ]; laws/policies [ ]; causal claims [ ].

Scenario transfer worksheet for AI Hallucination Check: current task → [ ]; reader/genre → [ ]; constraint → [ ]; tempting default → [ ]; better decision → [ ]; why it fits → [ ]; next related guide → [ ].

Skill progression worksheet for AI Hallucination Check: beginner version → [ ]; intermediate constraint → [ ]; advanced constraint → [ ]; what changed in the writing decision → [ ]; next skill to study → [ ].

Constraint decision sheet for AI Hallucination Check: normal case → [ ]; harder constraint → [ ]; what must stay true → [ ]; what can change → [ ]; evidence/reader/genre check → [ ]; specialist skill to study next → [ ].

Misconception correction audit for AI Hallucination Check: shortcut or assumption → [ ]; why it fails here → [ ]; corrective move → [ ]; principle that must remain true → [ ]; goal-based next skill → [ ].

Failure-mode and reader-intent audit for AI Hallucination Check: visible symptom → [ ]; likely underlying problem → [ ]; reader need → [ ]; competing technique considered → [ ]; correction → [ ]; evidence/genre/format constraint → [ ]; final verification → [ ].

Boundary-and-tool-choice audit for AI Hallucination Check: reader objection → [ ]; actual writing job → [ ]; evidence/genre/format boundary → [ ]; tempting alternative → [ ]; why it is the wrong tool here → [ ]; correct adjustment → [ ]; final verification → [ ].

Evidence/support and mastery ladder for AI Hallucination Check: minimum acceptable → [ ]; strong → [ ]; expert → [ ]; audience/genre constraint → [ ]; evidence or story support → [ ]; boundary that would require a different technique → [ ].

Certainty and expert-exception audit for AI Hallucination Check: statement or choice → [ ]; evidence/canon/authority supporting it → [ ]; confidence level → [ ]; uncertainty that must be named → [ ]; expert exception → [ ]; nearest alternative technique → [ ]; why the final choice fits better → [ ].

Purpose-and-verification decision audit for AI Hallucination Check: real writing job → [ ]; plausible choice A → [ ]; plausible choice B → [ ]; downstream consequence of each → [ ]; facts/evidence/canon to verify → [ ]; reader or genre boundary → [ ]; final purpose-fit reason → [ ].

Publication-readiness audit for AI Hallucination Check: claim/advice/example → [ ]; strongest available source tier → [ ]; important disagreement or uncertainty → [ ]; what would change the conclusion → [ ]; publish now / qualify / revise / hold → [ ]; final verifier or source of truth → [ ].

Provenance and classification audit for AI Hallucination Check: draft statement/example → [ ]; classify as fact / interpretation / recommendation / convention / original example → [ ]; source or internal provenance → [ ]; source date / version / checked date → [ ]; attribution needed → [ ]; recency risk → [ ]; final publishable wording → [ ].

Source-record and update audit for AI Hallucination Check: statement / example → [ ]; external fact, project fact, interpretation, recommendation, convention, or original example → [ ]; controlling first-party / primary source → [ ]; secondary source if used → [ ]; source record / locator → [ ]; date or version checked → [ ]; update trigger → [ ]; publish / qualify / revise / hold → [ ].

Source-ledger and revision-history audit for AI Hallucination Check: claim / recommendation / example → [ ]; source or project record ID → [ ]; source status active / superseded / corrected / retracted / retired → [ ]; claim-level last verified date/version → [ ]; event that triggered recheck → [ ]; correction / withdrawal action → [ ]; reader-facing disclosure needed? → [ ]; what changed and why → [ ]; prior wording/version retained at → [ ].

Claim/source status audit: Item [claim/source/example] | Status [active / review due / stale / corrected / retracted / superseded] | Controlling source/version [x] | Last verified [date/version] | Stale trigger [x] | Correction severity [0–3] | Reader disclosure [none / note / correction / withdrawal] | Changelog entry [what changed + why].

Dependency and propagation audit: Claim / example / recommendation [x] | Controlling source / project record [x] | Dependency strength [direct / shared / advisory / independent] | Review trigger / reason code [x] | Needs-review scope [local / direct dependents / cluster] | Replacement source requirements [x] | Carry-forward decision [x] | Propagation targets [x] | Reader disclosure / changelog action [x].

Persistent editorial registry audit: Topic slug → [ ]; claim UID/key → [ ]; source UID/key → [ ]; dependency UID + strength → [ ]; source status/version → [ ]; open review UID/reason/priority → [ ]; replacement source if any → [ ]; reviewer + resolution → [ ].