Why Generative AI May Be Making Your Managers Worse, Not Better

A new argument is emerging in AI research circles: generative AI may be making managers less effective, not more. While that sounds counterintuitive given the productivity promises, the reasoning is worth taking seriously.

The Core Problem

Management effectiveness relies heavily on soft skills—knowing when to ask questions, how to read a room, when to push back, and how to coach someone through ambiguity. These skills develop through experience: through friction, uncertainty, and the discomfort of not knowing the answer.

Generative AI short-circuits that process. When a manager can instantly generate a performance review, a meeting summary, or a decision memo, they lose the deliberate practice that builds judgment over time. The output looks polished. The skill development doesn’t happen.

What the Research Is Flagging

Legal scholars are beginning to document this pattern in professional settings. In law, junior associates who rely on AI-generated briefs may never develop the reasoning skills required for complex litigation. The same logic applies to management: leaders who outsource cognitive work to AI tools may find themselves less capable of doing that work independently when it matters most.

Key Takeaways

• AI tools that replace cognitive effort—rather than augmenting it—may slow the development of managerial judgment.
• Friction in management tasks (writing a difficult email, thinking through a promotion decision) isn’t inefficiency. It’s where skill development happens.
• Organizations should distinguish between AI that automates low-value tasks and AI that removes the deliberate practice behind high-value ones.

This doesn’t mean managers should stop using AI tools. It means they need to be intentional about which tasks they delegate—and which ones they keep doing themselves, even when AI could do them faster.

🔗 Read the full article on The National Law Review