Publications

Why Leaders and Executives Shouldn’t Be Afraid of AI

By Natalie Gill-Mensah · July 2026

The leaders and executives I interact with are rarely afraid of AI the way the headlines imagine. No one is lying awake picturing a robot in their chair. The fear is quieter, and honestly, more reasonable than that. It sounds like this: I don’t understand this well enough yet, it’s moving faster than I can follow, and I’m the one who has to answer for it.

Call that what it is. A leadership instinct. Every executive who has been around a while has watched a shiny technology roll in on a wave of promises and roll back out having burned a budget and a few good people. Skepticism you earned that way is worth keeping.

So I don’t ask leaders to drop their caution. I ask them to aim it somewhere more useful. What you’re after is governed confidence, the discipline of moving while you actively manage what could go wrong. Fear and reckless adoption are two ways of failing at the same job, and the ground between them is exactly where good executives already live.

Left alone, though, the fear costs more than it protects. Here’s where.

You can’t afford to stand still

The bar quietly moved. Your board, your clients, and your staff now assume you’ll keep pace with tools that make everyone around you faster. Your competitors already have. They aren’t holding meetings about whether to use AI. They’re banking small gains every week: the research that used to eat an afternoon, a first draft in minutes, the report read and summarized before the meeting even starts.

The threat rarely arrives as the dramatic disruption the think-pieces promise. It shows up as erosion. A competitor who turns proposals around in half your time. A five-person shop that produces like a team of fifteen. You don’t watch it happen. You notice it in the bid you lost and the gap you can’t quite explain. Standing still feels safe because it feels familiar, but it’s still a decision, and it bills you quietly, every month.

The real risk isn’t AI. It’s ungoverned AI

The reframe that lands hardest in a room is this: if you’re worried AI is loose in your organization, I have bad news. It already is. Your people are using it right now. Pasting documents into free tools over lunch. Cleaning up client emails. Summarizing the report they didn’t have time to read. They’re doing it with or without your blessing, and usually without mentioning it.

That’s where your exposure actually sits. Technology used in the dark, with no rule about what data goes where and no one checking what comes back out. A ban only blinds you to it. The instinct to prohibit and wait produces the least safe outcome on the table. Your job is to bring the practice into the light, where you can shape it.

The guardrails already exist

Confidence is available to you because responsible adoption isn’t uncharted. The controls are ordinary, and most of them are disciplines you already run for other risks.

A working setup needs only a few things. A clear line about what information never goes into a public tool. A human reading anything before it reaches a client or the public. The habit of vetting a tool before you adopt it, not after it’s already living in twelve workflows. Honesty about when AI was used. And one page, written down, that turns all of it from hallway assumption into a shared rule. Every one of those is a judgment call, which happens to be what you’re already paid for. No data scientist required.

You won’t lose your voice, unless you hand it over

Leaders and communicators tend to name a subtler fear than job loss. They worry about sounding like everyone else. About feeding a hard-won voice into a machine and getting back the same flat, agreeable tone we can all now spot from across the room.

It’s a fair worry. What I’ve found in practice runs the other way. A while back I had to give remarks on short notice, and a colleague nudged me to draft them with AI. I was skeptical, so I didn’t ask it for a speech. I gave it my outline and told it exactly how I wanted the room to feel. It handed back a scaffold. Then I did the part that mattered: I found the quotes myself, the ones that carried my actual values, and I wove them in. What I delivered was mine. The tool just cleared the path to it faster.

That’s the whole distinction. Leaders who come out sounding generic are the ones who shipped the first draft. The ones who keep their voice let AI take the underbrush, the research and the structure and the ninth revision, and stay the author of everything that counts. Judgment, taste, the point of view only you have. Those were never the machine’s to give. Used that way, AI gives your voice more room.

Adoption is a people problem before it’s a technology problem

This is the part most AI conversations skip. In twenty years of moving organizations through change, across international development and the private sector, the hard part was always the people.

Change fails for human reasons. A mandate no one understood. A rollout with no owner. Staff who were told what was changing and never why, so they nodded in the meeting and quietly kept doing it the old way. AI lands the same. The organizations that stumble usually chose a perfectly good tool, then ran a human transition like a software install.

I’ve watched it cut the other way too. Give a team the chance to see their own words reflected back, what they actually said they needed, organized and surfaced so they can build on it, and buy-in arrives on its own, because they own it. That’s leadership work. If you’ve carried people through hard change before, you already hold the scarcest skill this moment asks for. That experience is your edge.

What to do this week

Confidence sticks better when it’s attached to something you did. If any of this lands, start small and start now.

Find out how your team is already using AI. Ask plainly, promise no penalty, and assume the honest answer is “more than you’d guess.” Write one page: what’s encouraged, what’s off-limits, what data must never leave the building. Pick a single workflow that eats real time and run a contained pilot, one team, one process, a clear before and after. Then put someone’s name on it, because adoption without an owner drifts.

None of that is dramatic. That’s the point. The executives who do best here will be the ones who took an honest, well-earned caution and turned it into governance. That turn is the work. It always was.