The Leadership Friction  |  Edition No. 29  |  September 15, 2026

The Human Challenge in an AI World  |  Part 5 of 8

AI Can Give You an Answer. Leadership Still Requires Judgment.

A fast answer is not the same thing as a sound decision.

AI can hold more information than any leader can carry in their head, but that doesn’t make its recommendation correct. This edition is about the judgment leaders still owe their organizations, even when the answer arrives already reasoned out.

Ask AI a question, and you’ll have an answer in ten seconds. Well-organized. Confidently worded. Dressed up with next steps.

That speed creates a temptation, and I am watching leaders fall into it right now. They start mistaking a fast answer for a sound decision.

Here’s what I see: those are not the same thing. One is information. The other is judgment. And a leadership team that stops telling the difference has a bigger challenge coming than a slow decision cycle.

The Strength

What AI Actually Does Well

AI can hold more information than any one leader could carry in their head. It reads through more material than you’d get through in a week, notices patterns buried in the numbers, compares several scenarios side by side, summarizes a document nobody had time to read closely, and hands you three or four courses of action already reasoned out.

That’s useful. I’m not arguing otherwise.

But useful isn’t the same as right. None of what I just described makes the final call correct. It makes the final call easier to arrive at. Those are different things, and the gap between them is exactly where leaders are getting into trouble.

Here’s what that looked like in practice, recently. I’m doing some work right now with a client who’s building out a new sales team. She sent me a stack of documents to work from, and I ran them through my AI tool and asked for a summary. It gave me one, then mentioned, almost in passing, that it had already made several changes based on its own logic and its own recommendations. When I asked it to redo the summary, it stopped and asked me to accept or deny those changes first.

It had already exercised judgment on her documents. I just hadn’t been told until I pushed back.

The Gap

Judgment Requires Context AI Doesn’t Have

A recommendation can look complete on paper and still be wrong, not because of what’s in it, but because of what’s around it: organizational history, timing, how ready the team actually is, a customer relationship that took years to build, the cultural cost, how much trust is already in the room, how much risk the organization can absorb right now, or a piece of information the system was simply never given.

AI can evaluate the information it has. Leaders are responsible for recognizing what the information does not tell them.

That’s the job. It always has been. AI just made it easier to skip.

The Risk

The Danger of Borrowed Confidence

This is where it gets dangerous. AI communicates with confidence whether or not the answer deserves it. The tone doesn’t change based on how solid the reasoning is underneath.

The challenge is that confidence and correctness feel identical from the outside. So leaders stop asking the questions they used to ask automatically.

• What are we missing here?
• What assumptions is this built on?
• Who gets affected if we’re wrong?
• Does this actually fit what we know about our own organization?

The output starts to feel authoritative simply because it sounds authoritative. That’s borrowed confidence, and it costs more than it looks like it does.

The Responsibility

Judgment Cannot Be Delegated

AI can support a decision. It can challenge your thinking. It can generate options you hadn’t considered and surface a risk you missed. What it cannot do is own the outcome.

Someone still has to make the call and stand behind it when it doesn’t go the way the recommendation promised. We talked about ownership in Edition 27, and this is the same thread pulled tighter. This is a people question before it’s a technology question. AI can change the systems around a decision. It cannot change who is accountable for it.

The Practice

What Good AI-Assisted Judgment Looks Like

Before acting on a recommendation, I’d want a leader working through four questions.

• What does the data say, and what is this recommendation based on?
• What does the data not say, meaning what context or history might be missing from it?
• Who is affected, both operationally and personally?
• What assumptions are we making?

Then comes the question that matters most: who owns this? Someone has to be able to answer that before anyone acts.

The Shift

The Standard Is Going Up, Not Down

One thing I’ve learned watching organizations adopt AI: as it takes over more of the routine analysis, leaders need sharper judgment, not less of it. The value of leadership keeps shifting toward interpretation, discernment, and the willingness to own a consequence.

That’s the real counterpoint to the idea that AI simply makes leadership easier. It doesn’t. It raises the bar on exactly the parts of the job that were never going to be automated in the first place.

The next time AI hands you a recommendation, don’t just ask, “Is this a good answer?” Ask, “What would I need to understand before I’m willing to own this decision?”