---
id: "action-encourage-second-guessing"
type: "action-item"
source_timestamps: ["§ How to Use Explainable AI Responsibly"]
tags: ["leadership", "culture"]
related: ["concept-algorithmic-override", "quote-stop-asking-why", "framework-responsible-xai-deployment"]
action: "Cultivate a management culture that explicitly encourages and rewards employees for second-guessing AI recommendations."
outcome: "Preserves human judgment and increases the rate of justified algorithmic overrides."
speakers: ["Alex Chan"]
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"
---
# Actively encourage employees to second-guess AI

**Action:** Cultivate a management culture that explicitly encourages and rewards employees for second-guessing AI recommendations.

**Expected outcome:** Preserves human judgment and increases the rate of justified [[concept-algorithmic-override|algorithmic overrides]].

Managers need to foster a culture where **challenging the algorithm is viewed as a positive demonstration of human judgment**, rather than inefficiency or insubordination. This combats the long-term risk of employees losing their critical-thinking skills (see [[quote-stop-asking-why]]) and directly leverages the finding that engagement raises override rates ([[claim-explanations-increase-override]]). It is prong 3 ("value human judgment") of the [[framework-responsible-xai-deployment]].

**Enrichment note:** Complements the automation-bias and human-in-the-loop literatures, which stress that explanations are effective only when integrated into workflows with accountability. **Nuance:** apply proportionately — high-stakes decisions warrant enforced second-guessing/documentation, while routine ones may not, to avoid alert fatigue (see [[question-ui-ux-for-forced-engagement]]).


## Related across articles
- [[action-celebrate-error-catching]]