How Do Leaders Keep Their Own Point of View When Using AI?

Leaders keep their point of view when using AI by treating the decision as theirs to own, not the model's to make. AI hands you more options and reflects your own thinking back at you. That deepens indecision more often than it resolves it. The anchor is internal. Decide what you're for, then commit and move.

Anchored agency erodes as a pattern, not a moment

It rarely shows up as a single bad call. It shows up as treading water.

A leader knows there's a decision on their desk. They circle it. They talk about it session after session. They know there's a problem and they know they have to move. Instead they redirect. They go solve a problem they already know how to solve, because the one that matters is the one that scares them.

This isn't laziness. It's fear, and the fear is earned. Early in a career, decisions are cheap. You swing, it works out, you move on. Then you climb, and a lot more hangs in the balance. Now the call carries political fallout, cultural fallout, career risk. So the leader freezes. It looks like stuck.

And it turns out the freeze is the norm, not the exception. McKinsey surveyed more than 1,200 executives and found that fewer than half say their decisions are timely, and 61 percent say most of the time they spend making them is ineffective. The cost isn't abstract. For a typical Fortune 500 company, that indecision burns more than 500,000 working days a year — roughly $250 million in wages spent deciding badly or not at all. That's the number to put in front of a CFO who thinks leadership development is soft.

Because the person who pays for the freeze first is rarely the leader. It's the person underneath them trying to execute. She brings a proposal. He says it's a great idea and signs on for a week. She comes back with the next step ready to go, and he gets scared, and the can gets kicked down the road again. She's left holding a job she can't do, because the job requires a decision that never comes. Just stand where you are and make a decision so I can do the thing you hired me to do.

AI didn't cause the freeze. It made it deeper.

The freeze predates AI. What AI did was pour an accelerant on it.

Now the frozen leader has somewhere new to go. They ask the model what they should do. It presents them with more options, often the same options they already had, and now a machine is saying it back to them. They're not closer to a decision. They're more paralyzed than before.

The research backs the pattern. A field experiment by Leonard Boussioux and colleagues at Harvard Business School and MIT Solve found that when AI gets convenient enough, people start deferring to it without real scrutiny — and that the most effective setup is the one where the human still makes the final call and lets AI guide the thinking, not the reverse. The danger isn't that the model is wrong. It's that its answers arrive polished and confident, and a leader already looking for a reason not to decide takes the reflection as a verdict.

It works like standing at the edge of the surf. The tide pulls the sand out from under your feet, then it refills, then it pulls again, and if you stand there long enough the current decides for you and moves you somewhere you didn't choose to go. The social forces were already thick enough to do that. AI is a second current layered on top of the first, and it makes the pull exponentially harder to resist.

There's a quieter cost too. The longer you let the mirror set your direction, the more your direction looks like everyone else's, because everyone else is feeding the same prompts to the same model. The leader stops being the source of the point of view.

Nobody wants to own the call — which is exactly why you have to

Here's the part that doesn't get said out loud. Almost no one in the room wants to own the decision.

The coach won't tell you what to do, because that isn't the job — the job is to help you find it yourself. The consultant hedges, because if the call goes wrong the liability is real and the brand has to be protected. AI makes a bad call and simply asks what's next. It doesn't lie awake. It doesn't lay anyone off. It doesn't go broke. Nobody wants to stand up and say this is what needs to happen, because the risk of being wrong is real and it lands on a human being.

But somebody has to grow roots. Somebody has to grow vision. Somebody has to grow the anchored agency to plant a flag and stay next to it when the wind comes. That vacuum isn't a reason to wait for more certainty. The vacuum is the job.

So here's what I'd tell the leader doing the punting. Not choosing is a choice, and it's usually the worse one. You are the author of this, not a victim of it. You won't get every call right, and that was never the standard. But a leader who runs the work — asks the sharper question, judges what the data leaves out, then commits — bats far stronger than the leader who hedges from zero conviction and never takes the swing.

Anchored agency is one of the five capacities AI can't replace — the through-line of what AI can't replace in leadership. The leaders who keep their point of view in this era won't be the ones who learned the tools fastest. They'll be the ones who stayed the author.

If you're the one trying to execute beneath a leader who keeps punting the call, this is the language for the conversation you've been having in your own head.