Ford CEO Jim Farley has said artificial intelligence could replace half of America’s white-collar workforce. Microsoft AI CEO Mustafa Suleyman expects most white-collar tasks to become automatable within 12 to 18 months. Anthropic CEO Dario Amodei has warned that half of entry-level office jobs could disappear within one to five years.
54% of layoff events tracked in 2026 have cited AI, automation, or machine learning as a factor, compared with fewer than 8% in 2025, according to IBT.
At the same time, employees inside many companies are hearing a much softer message. AI will remove busywork. It will make people more productive. It will help teams do more.
I have heard both versions of this conversation.
Technology has always changed workforce needs. What concerns me is what happens when leadership has a clearer view of where work is going than employees do and chooses not to share it.
That is where an AI transformation becomes a leadership issue.
The Two Rooms Problem
Behind closed doors, executive AI conversations tend to revolve around a very different set of questions.
How much capacity can AI absorb? Can we grow without adding headcount? Which positions no longer need to be backfilled? How does automation improve margins?
Employees hear a different vocabulary. AI will empower you. AI will eliminate repetitive work. AI will make you more productive.
Both narratives can technically be true. They describe very different futures.
Employees are also far more aware of that difference than some leaders assume. They listen to earnings calls, read investor presentations, notice hiring freezes, and watch positions remain unfilled. They use the same AI tools their companies are buying, so they can see which parts of their own work those tools can already absorb.
More and more, employees are conducting due diligence on their own employer.
I made a related point in Wall Street Just Walked Into the AI Lab. Once a company starts explaining its AI strategy to investors in terms of productivity, margins, and headcount efficiency, that information does not stay inside the investor presentation. Employees can compare it with what they hear at the town hall.
That is why “AI won’t take your job” is one of the worst promises a CEO can make today.
Leaders cannot guarantee jobs they may later need to change or eliminate. False reassurance creates trust debt, and that debt comes due when the workforce eventually changes.
What a Company Owes Employees Before Their Work Disappears
We already have severance. It comes after a job disappears.
AI raises a different question. What does a company owe employees before their work disappears?
My answer is strategic notice.
Employees need enough visibility into where the organization believes work and value are moving to make informed choices about their own future.
Strategic notice does not require a CEO to disclose confidential restructuring plans. It requires honesty about direction.
A credible CEO should be able to say:
“We expect AI to absorb much of this work. We do not know exactly what that means for headcount yet. Here is what we are learning, and here are the capabilities we believe will matter.”
That conversation gives employees something useful: information they can act on.
If leadership believes there is a 60% probability that your role will not exist in two years, when do you deserve to know?
Leadership doesn’t owe employees certainty. It owes them visibility.
Don’t guarantee my job. Guarantee I won’t be the last to know it’s disappearing.
The Biggest AI Layoff May Be the Person You Never Hire
Much of AI-driven workforce reduction may never appear in a mass layoff announcement.
It will happen quietly.
A company hires fewer entry-level employees. A vacancy never gets backfilled. A team of ten becomes a team of seven and handles the same volume. A new function that once would have required twenty people launches with eight.
Nobody receives a termination notice because many of those jobs are never created.
That could be more consequential than the headline layoff numbers.
It also creates another workforce problem. Entry-level work has always done more than produce output.
Research, first drafts, basic analysis, document review, and preparation were how people learned judgment. They looked like routine tasks, but they also functioned as apprenticeships.
If AI removes those tasks, organizations have to redesign how people develop expertise.
That concern connects directly to what I wrote in We’re Raising a Generation That Outsources Thinking. Removing routine cognitive work creates immediate efficiency. The longer term question is whether we also remove the experiences that teach people how to think through unfamiliar problems on their own.
What happens when we automate the first five rungs of the career ladder?
The biggest AI layoff will be the person a company never hires.
When Change Management Becomes an Integrity Problem
Change management should help employees navigate uncertainty.
It becomes a leadership integrity problem when its purpose shifts toward shielding employees from uncertainty that leadership is already planning around.
A credible CEO should be able to say, “We do not know exactly what happens to every job. We do know which work is becoming automatable, and we know where human value is moving.”
I would rather hear a CEO say, “I don’t know, but here’s what I know today,” than offer reassurance nobody can responsibly guarantee.
There is a simple test for this.
Put the investor call transcript beside the employee town hall transcript.
Do they describe the same AI future?
If your AI strategy sounds different on the earnings call than at the employee town hall, employees will believe the earnings call.
Change management becomes manipulation when its purpose is keeping people comfortable rather than keeping them informed.
The damage goes beyond attrition. Poorly handled AI adoption can erode trust, initiative, and dignity even when employees technically comply with the rollout. I explored that problem in The Hidden Cost of AI Adoption: Lost Dignity. A workforce can follow instructions while quietly withdrawing the judgment, creativity, and discretionary effort the company still needs from them.
Employees Have Responsibility Too
Employees should not wait for management to announce that their work is changing.
“Learn AI” is insufficient advice.
A better question is: Where is value moving as AI makes parts of my current expertise abundant?
If information becomes cheap, judgment becomes more valuable.
If execution becomes cheap, problem selection becomes more valuable.
If analysis becomes abundant, knowing what deserves analysis becomes more valuable.
The goal is to identify tomorrow’s source of human value.
That is also the deeper argument behind AI Didn’t Take Your Job. Someone Using AI Did.. Protecting a current job description is a losing strategy when technology is changing the work inside that description. The more useful response is understanding which parts of your contribution remain scarce as other capabilities become abundant.
Companies have a responsibility here too because leadership usually has access to information employees do not.
Executives do not need to predict every job that will exist two years from now. They do need to give people enough truth to prepare for where the organization is heading.
That is the standard I would apply to any AI transformation.
Leadership does not owe employees certainty about the future. It owes them enough visibility to make informed decisions about their own.
If people are going to lose parts of their work, change careers or develop entirely new capabilities because of your AI strategy, they should not learn about that strategy after the market already has.
