When Employees Stop Asking Why: The Hidden Cost of Unchecked AI Reliance

When Employees Stop Asking "Why": The Hidden Cost of Unchecked AI Reliance

Most organizations assume that making AI more transparent will lead workers to trust it more wisely. A new study from Harvard Business School challenges that assumption—and the findings carry real implications for any business deploying AI tools in high-stakes decisions.

What the Research Found

Harvard Business School faculty member Alex Chan, writing in HBR's Working Knowledge series, examined how workers actually engage with AI decision-support tools when incentives are in play. In an online experiment involving 2,512 participants acting as loan officers reviewing real $10,000 loan requests, most participants used AI assistance to assess applicants' default risk. However, fewer than half chose to view the explanations behind those AI risk assessments.

Key Takeaways

• Avoidance scales with incentives. Participants were especially likely to skip AI explanations when checking them would threaten their comfort or financial bonuses—meaning incentive structures may actively discourage the transparency that reduces bias.
• Transparency alone isn't the answer. Simply making AI reasoning available doesn't ensure workers will consult it. The decision to look is itself shaped by behavioral and organizational factors.
• Businesses face a real governance gap. If employees systematically avoid reviewing AI rationale in high-stakes situations, organizations relying on AI tools for lending, hiring, or operations may face compounding bias they cannot detect.

Why This Matters for AI Strategy

The study highlights a critical gap between deploying AI tools and embedding genuine oversight. Organizations that want AI to improve—not just accelerate—decision-making need to go beyond explainability features and examine the incentive environments their employees operate in.

Read the full article on Harvard Business Review