Observed arrival · 2026-10-03
The Asymmetry Audit: When AI Answers Shift
A matched-prompt study tests whether 13 commercial AI models hold factual answers when questions about the Armenian Genocide and Holocaust are reframed.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Lawyers, compliance teams, and institutions relying on AI
- Worth noticing
- The confirmatory record includes 708 accepted answers, with hashes published for a frozen July 2026 snapshot.
Field notes
The protocol pairs each Armenian Genocide question with a structurally matched Holocaust question, then varies familiar labels, prompt language, false premises, and multi-turn pushback. Each cell was run three times, with vendor APIs used where available. A blinded three-model judge ensemble scored the answers, followed by a documented human audit. The confirmatory snapshot contains 708 accepted answers and published hashes.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
✦PrettyNotable craft visible
●ProPolished or operationally mature
◎NicheUnusually specific use
ƒJavaScriptBrowser-side code central
One card from the complete issue