AI Isn't Replacing Workers. Leadership Is.

A few different countries have started saying the quiet part out loud: AI is going to replace large portions of the workforce, and they're not going to pretend otherwise.

The rest of the business world is paying attention. Boardrooms are running the numbers. Consultants are pitching "AI workforce optimization" decks that are really just headcount-reduction plans with a new label. And the conversation has clearly moved from "if" to "when."

I'm not going to pretend it isn't happening. The technology can do it. The economics encourage it. And the path of least resistance is already being walked by companies who'd rather cut than rebuild.

But "can" and "should" are different conversations. And the ethical question that everyone is rushing past is the one that actually matters.

Tasks vs. humans

Most of the headlines collapse two very different things into one sentence.

AI is brilliant at replacing tasks. It's mediocre at replacing humans.

Tasks are repeatable, codifiable, pattern-based. They can be described in a prompt, modeled from a few examples, and reproduced reliably enough that automating them is genuinely valuable. AI does this better than us, more consistently, and around the clock.

Humans are not tasks. A person in a role brings judgment, context, history with the clients, relationships with the team, taste about what "good" looks like, and the ability to recognize when something is wrong even though every metric says it's fine. They can hold ambiguity. They can read a room. They can be trusted with information that doesn't have a process around it yet.

When a company says "AI replaced 30% of our team," what they usually mean is: "AI absorbed 30% of the tasks, and we chose to cut headcount instead of redeploying the people who were doing them."

Those are not the same statement. The first is a description of technology. The second is a description of leadership.

The leadership decision

Once you separate the technology question from the leadership question, the ethics get clearer.

Option A: Use AI to compress the cost of every existing process and cut the people whose work has been compressed. Hit the quarter. Save the line item. Move on.

Option B: Use AI to free people from the work that didn't actually need them, and redeploy them into the work that only humans can do. Spend the savings on reskilling, not severance.

Both options are legal. Both have been chosen. The first is faster and cheaper in the short term. The second is harder, slower, and produces a wildly different company at the end.

The ethical question isn't whether AI should be used. It's whether the humans whose roles change deserve a real path forward, or just a polite goodbye.

I don't think that's actually a hard question. It's an inconvenient one.

What "redeploy" actually means

It's easy to say "we'll just reskill the team." It's harder to do it in a way that respects the people involved.

Reskilling isn't a slide in the all-hands deck. It's not a Coursera license. Real redeployment looks like this:

A clear new role description, written before the old role is dissolved. Not a vague "we'll figure something out."

Time and structured learning, paid, with measurable milestones. People can't learn AI tools in the cracks between their existing job and their commute.

A leader who actually believes the person can grow into the next thing — and tells them that out loud. The most overlooked variable in any career transition is whether the person above you thinks you can do it.

A clear runway. Six months, not six weeks. The honest truth is that not every transition will work. But people deserve the chance to find out.

And the part most companies skip: a real assessment of which roles are growing because of AI, not just which ones are shrinking. The companies that will win this decade are the ones who staff up on judgment, taste, and customer-facing work — not the ones who just cut and call it transformation.

The ten-year view

Here's why this matters beyond the moral argument.

The companies that cut hardest in the next two years will look brilliant on a quarterly call and average-to-bad on a decade timeline. They'll have leaner P&Ls and emptier benches. When the next thing emerges — and there is always a next thing — they'll have no one with enough institutional context to recognize it, and no one with the loyalty to push through it.

Meanwhile, the companies that took the harder path will have a workforce that grew alongside the technology. People who understand both the AI and the business. People who can build the next thing because they understand the last thing.

That's not a soft argument. That's a competitive one.

The market rewards the version of operations that compounds. Cutting decompounds. Building compounds.

Where leadership comes in

I keep saying "leadership decision" because that's exactly what it is. The technology is going to keep advancing whether any one company chooses well or badly. The countries setting precedent on workforce displacement are going to keep setting precedent. None of that is up to us.

What is up to us — what is up to anyone running a team, anyone in an operations seat, anyone with a vote in how this rolls out — is what we owe the people whose work is being changed.

I don't think the answer is to slow the technology down. I'm using it every day. I'm teaching my team to use it. I believe in what it can unlock.

But I'm also clear-eyed about the choice in front of the people running companies. The technology hands you a lever. It doesn't tell you which direction to pull it.

The people who pull it toward "cut" will get a short-term win and a long-term mess. The people who pull it toward "build" will get a workforce that can carry them into whatever comes next.

That's the ethics. That's the strategy. They're actually the same conversation.