---
id: "entity-preference-for-explanations-paper"
type: "entity"
entityType: "other"
canonicalName: "Preference for Explanations: Case of Explainable AI"
aliases: ["HBS Working Paper 26-028", "NBER Working Paper 35240"]
source_timestamps: ["¶7"]
tags: ["research-paper", "literature"]
related: ["entity-alex-chan", "concept-willful-ignorance-in-ai"]
sources: ["adoption"]
sourceVaultSlug: "hbr-seg-adoption"
originDay: 9
articleStem: "hbr-edu-37-employees-not-questioning-ai"
sourceUrl: "https://hbr.org/2026/06/employees-arent-questioning-ai-advice-enough"
sourceTitle: "Employees Aren’t Questioning AI Advice Enough"
---
# Preference for Explanations: Case of Explainable AI (Working Paper)

**Type:** Other (academic working paper / publication) · **Canonical name:** *Preference for Explanations: Case of Explainable AI*

A working paper by [[entity-alex-chan|Alex Chan]], updated **February 2026**, which challenges the assumption that people naturally want more transparency from AI systems. It details the loan-approval experiment in which **2,512 participants act as loan officers** using an AI default-risk predictor, choosing whether to view explanations and deciding loan approvals. It is the central empirical basis for every claim in this vault (see [[concept-willful-ignorance-in-ai]], [[concept-moral-quandary-avoidance]], [[concept-algorithmic-override]]).

**Canonical references (enrichment):**
- HBS working paper page: `hbs.edu/faculty/Pages/item.aspx?num=68104`
- NBER working paper: `nber.org/papers/w35240` (HBS WP 26-028 / NBER WP 35240)
- SSRN manuscript: `papers.ssrn.com/sol3/papers.cfm?abstract_id=6824998`
- Related summary article by Chan: *"Explanations on Mute: Why We Turn Away From Explainable AI."*

**Verification caveat:** Several precise magnitudes cited in this vault (nearly 20% more declines under bonus pay; +10pp / 23% avoidance under bias salience; ~6pp higher override) are **not independently visible in public summaries** and should be confirmed against the full paper tables before being quoted as settled figures.