Skip to content

R0044/2026-04-01/Q003/H1

Research R0044 — Expanded Vocabulary Research
Run 2026-04-01
Query Q003
Hypothesis H1

Statement

Explicit, formal vocabulary bridging exists between human-factors automation bias/overtrust concepts and AI safety sycophancy concepts, published in regulated-industry research or guidance.

Status

Current: Eliminated

Supporting Evidence

Evidence Summary
None No source provides a formal, deliberate vocabulary mapping

Contradicting Evidence

Evidence Summary
SRC01-E01 Ibrahim et al. discuss both concepts but do not provide a formal vocabulary mapping
SRC03-E01 Malmqvist treats sycophancy as a purely technical LLM problem without connecting to human factors

Reasoning

No publication was found that explicitly states "automation bias and sycophancy describe the same phenomenon" or provides a formal cross-domain vocabulary mapping. H1 is eliminated.

Relationship to Other Hypotheses

H2 is better supported — partial bridging exists. H3 is too strong given Ibrahim et al.'s work.