Skip to content

R0044/2026-04-01/Q003/SRC01/E01

Research R0044 — Expanded Vocabulary Research
Run 2026-04-01
Query Q003
Source SRC01
Evidence SRC01-E01
Type Analytical

Ibrahim et al. unified framework bridging cognitive science, HCI, and AI safety concepts

URL: https://arxiv.org/html/2509.08010v1

Extract

Ibrahim et al. provide the most comprehensive vocabulary bridging found in this research:

Concepts explicitly connected in a single framework: - Automation bias (cognitive science / human factors) — listed as a cognitive mechanism behind overreliance - Sycophancy (AI safety) — described as models' "tendency to be overly agreeable" that "discourages critical evaluation" - Trust calibration (HCI / human factors) — extensive discussion of miscalibration where users "struggle to generalize LLM successes and failures across domains" - Cognitive offloading (cognitive science) — mechanism by which users delegate cognitive work to AI - Complacency bias (human factors) — listed alongside automation bias

System-side design recommendations that bridge both traditions: - Uncertainty expression: models should communicate limitations through "natural language (e.g., 'I'm not sure, maybe...') or other forms (e.g., confidence scores)" - Friction mechanisms: "Extra click or disclaimers" to slow users at critical points - Mixed-initiative controls: designs where "both the user and AI suggest, produce, evaluate, and modify outputs" - Visual cue adjustment: revising features like typing indicators that "contribute to the illusion of human-like behavior"

Critical observation: While Ibrahim et al. use both vocabulary sets in a single paper, they do not explicitly declare the bridging as their purpose. The paper's focus is overreliance measurement and mitigation, and the vocabulary bridging is a byproduct of their comprehensive literature review rather than a deliberate contribution.

Relevance to Hypotheses

Hypothesis Relationship Strength
H1 Contradicts Not a formal vocabulary mapping; bridging is implicit
H2 Supports strongly Best example of partial bridging — both vocabularies used in single framework
H3 Contradicts strongly Demonstrates at least some researchers are connecting the traditions

Context

This paper is significant because it demonstrates that the bridging is possible and productive — but it also demonstrates that even the best bridging work treats it as incidental rather than central. The absence of a deliberate, formal vocabulary mapping remains a gap.