Authority Laundering and AI Myth Inflation: Comment-Level Artifacts, Transcript, and Analysis of a Grok 4.20 Claim (January 2026)

This dataset documents and analyzes a comment-level epistemic failure surrounding claims that “Grok 4.20” independently solved novel mathematical problems beyond human capability. The materials originate from a January 2026 YouTube video published by the Dukta Feelgood channel and the subsequent public comment exchange. The dataset includes: Screenshots of the video catalog and thumbnails illustrating high-volume AI-themed content A verbatim transcript of the focal video asserting a Grok mathematical breakthrough Dukta_Feelgood.md A three-turn public comment exchange in which verification is challenged and deflected via claims of privileged access A structured Grok model query evaluating the video’s claims and identifying exaggeration and misattribution grok_prompt.md A formal analytical paper (Authority Laundering and AI Myth Inflation: A Comment-Level Case Study) articulating the epistemic pattern Claim → Authority → Access → Collapse Rather than adjudicating the ultimate truth of any mathematical result, this dataset examines how unverifiable access claims are substituted for evidence in platform-mediated AI discourse. The materials provide a reproducible micro-model of authority laundering, observer-induced dismissal, and narrative collapse under minimal epistemic scrutiny.