What Should I Tell My Team About AI and Their Jobs?

Share This Post
What Should I Tell My Team About AI and Their Jobs?

The question every founder is avoiding right now, and the specific, honest answer that keeps your best people from writing their own ending.


Someone on my team asked me, more or less directly, whether AI was going to take her job.

I gave her a good answer. Warm, reasonable, forward-looking. Something about how AI handles the boring parts and frees us up for the work that matters. She nodded. We moved on to the next agenda item.

It was a terrible answer. Not because it was false. Because it was vague, and vague is what people hear when you are hiding something. I know this because I watched her over the following weeks, and she was not reassured. She was polite. There is a difference, and I missed it for longer than I want to admit.

So here is the direct answer to the question in the headline. Tell your team the truth at the level of specific tasks, not job titles, and commit to something they can hold you to. Not “AI won’t replace you.” Not “we’re all going to have to adapt.” Say which tasks are changing, on what timeline, what happens to the person whose day is made of those tasks, and what you will do if you are wrong. Specificity is the entire message. Everything else is noise your team has already learned to discount.

Here is why this matters more than any tool you roll out this quarter. Your people are not waiting for you. Two loud, incompatible stories about AI and work are circulating right now, and your team has read both. One says the layoff spiral is a choice we are making. The other says the jobs apocalypse has begun and you had better prepare. Same technology. Same week. Two different futures.

When you say nothing, you do not stay neutral. You hand the microphone to whichever story is scarier, because fear is stickier than nuance. The vacuum always gets filled. It just does not get filled by you.

Key Takeaways

  • Silence is not neutral; when leaders say nothing about AI and jobs, employees default to the most frightening story available.
  • Only 13% of employees strongly agree that their organization’s leadership communicates effectively, according to Gallup, so your team already assumes you are not telling them everything.
  • Employee fear of AI-driven job loss jumped from 28% in 2024 to 40% in 2026 in Mercer’s survey of nearly 12,000 people worldwide.
  • The economic evidence genuinely points both ways, which means honest specificity beats confident prediction every time.
  • The highest-leverage move you can make this quarter is a conversation, not a tool rollout.

The Problem

Here is the trap I fell into, and I suspect you are in it too.

I did not stay quiet because I did not care. I stayed quiet because I did not know. I could not honestly promise anyone that their role would look the same in eighteen months. I have watched what these tools can do. I build with them every day. Promising stability I could not deliver felt like lying, and lying to your people is the one thing you do not get to take back.

So I did the thing that felt like integrity and was actually cowardice wearing integrity’s coat. I waited until I had a clear answer.

The clear answer never came. It was never going to come. And while I waited, the story got written anyway.

That is the part I want you to sit with, because it is the whole problem. You think your options are “reassure them” or “stay honest.” You think silence is the honest choice when you are uncertain. It is not. Silence is a message, and it is a loud one. Your team reads it. They read it in the pause before you answer, in the roadmap that quietly lost two headcount lines, in the fact that you have mentioned AI in every all-hands for a year and never once mentioned what it means for the person sitting in row three.

And they are not reading it in a vacuum. They are reading it against a backdrop where, in June 2026 alone, U.S. employers announced 14,029 job cuts with artificial intelligence cited as the reason, according to Challenger, Gray & Christmas. That was 31% of all cuts announced that month, and the fourth consecutive month AI led every other reason. Your team knows this. It is in their feed. It is in their group chat.

I once built a product I was proud of and watched it die in silence. A few hundred signups the first week, glowing messages, and six weeks later I could count the weekly active users on my hands. Not my fingers. My hands. What I learned from that is not “products fail.” What I learned is that the silence after the applause is information, and I was too slow to read it.

The silence in your company right now is information too. It is not peace.

Here is the reframe. You do not need certainty to speak. You need specificity. Those are different things, and confusing them is what keeps good leaders quiet.

The Evidence

I want to give you both sides of this honestly, because the honest picture is the argument.

The alarming case is real and it has math behind it. In June 2026, two economists, Brett Hemenway Falk at the University of Pennsylvania and Gerry Tsoukalas at Boston University, released a paper called “The AI Layoff Trap.” Their model describes a demand externality: each firm captures the full savings from automating a role, but bears only a fraction of the demand it destroys when that worker stops spending. The rest lands on competitors. Every firm behaving rationally produces a collectively irrational outcome, and the authors find that more competition and better AI make it worse, not better. They also find that most of the obvious fixes, including UBI, upskilling, and worker equity, do not resolve it in their model. Note that this is a working paper, not settled consensus, and it is a model rather than a measurement. But it is serious work from serious institutions, and dismissing it is not an option.

The reassuring case is also real and it has data behind it. The Budget Lab at Yale examined the U.S. labor market across the first 33 months after ChatGPT’s release and found no discernible economy-wide disruption in occupational mix or in employment by AI-exposure level. Brookings published findings under the plain title “New data show no AI jobs apocalypse, for now.” Read that last phrase carefully, because the researchers did. Their own caveat is that widespread effects would reasonably take longer than 33 months to appear.

And then there is the uncomfortable middle, which I think is closest to the truth. A Gartner study published in May 2026, based on 350 executives at companies with at least $1 billion in revenue that were actively deploying AI, found that 80% had reduced headcount. Here is the finding that should stop you: the companies that cut the most delivered nearly identical financial returns to the companies that cut the least, and in several cases the ones that cut less did better. Gartner’s Helen Poitevin put it plainly: chasing value only through headcount reduction is likely to lead most organizations down a path of limited returns.

Meanwhile, MIT Sloan professor emeritus Paul Osterman told Fortune in May 2026 that much of what we are watching is “AI washing,” companies using AI as a cover story for cuts they had already decided on. His line was blunt: “AI is a perfect excuse to justify big layoffs. It makes it seem as if it’s not our decision, our fault, it’s the technology.” Leaders have been promising smaller, leaner teams for twenty years. The label is what is new.

So: the doom case is a model, the calm case is a rearview mirror, and a meaningful share of the layoffs being blamed on AI may not be about AI at all.

Now look at what your people are carrying while the experts argue. Mercer’s Global Talent Trends 2026, drawing on nearly 12,000 executives, HR leaders, investors, and employees, found employee concern about AI-driven job loss rose from 28% in 2024 to 40% in 2026. The share of employees who say they are thriving at work fell from 66% in 2024 to 44%, lower than during the pandemic. And the finding I cannot get out of my head: 62% of employees say leaders underestimate AI’s emotional impact, while only 19% of HR leaders factor that impact into their digital implementation strategy.

Your team is not confused about the economics. They are confused about you.

The Solution: Name It, Map It, Commit It

You cannot resolve a debate that Wharton and Yale have not resolved. Stop trying. That is not your job.

Your job is much smaller and much harder. Your job is to tell twelve specific people what is true about their specific work at your specific company. That is a question you actually can answer, and you are the only person on earth who can.

Here is the structure I use now.

Name it. Say the thing out loud, including the part you are afraid of. “I know two stories are going around about AI and jobs. I know one of them is terrifying. I’ve been quiet about it, and I want to fix that.” You are not confessing weakness. You are demonstrating that you can hold a hard conversation without flinching, which is the only evidence anyone has that you will be honest with them later when it counts. If you have been silent, say you have been silent. They already know. Naming it costs you nothing and buys you everything.

Map it. This is where specificity lives. Jobs do not get automated. Tasks do. So go task by task, out loud, with the actual person. Which parts of this role are genuinely going to be done by a machine within a year? Which parts are not, and why? Which parts are you honestly unsure about? That third bucket is not a failure. It is the most credible thing you will say all day, because a leader who claims certainty about AI in 2026 is telling you exactly how much attention they have been paying.

Then say what happens to the person. Not the role. The person. If a third of someone’s week is about to be absorbed, tell them what fills the other third, or tell them you do not know yet and you are working on it with them rather than about them.

Commit it. End with something falsifiable. A commitment with a date, a number, or a name attached. “We are not reducing headcount because of AI this year, and if that changes you will hear it from me first and with sixty days’ notice.” Or “I am budgeting X hours a month for you to build AI skills on company time, starting Monday.” Or, if the honest answer is that this role is going to shrink, say that, and say what you will do about it, and say it early enough that the person has real options. That last one is the hardest sentence a founder ever says. It is also the one that people remember you for, in the good way.

I am not going to pretend I do everything right here. I ran my mouth about AI capability for a long time before I ever said a word about AI and my own team’s livelihoods. But I will tell you what changed when I finally had the conversation: nobody quit. Somebody cried. Somebody asked a question so good it changed our roadmap. And the tension I had been reading as “everything’s fine” turned out to have been something else entirely, sitting there the whole time.

People are not line items. Whatever you believe about where any of us came from, most of us agree that a human being is not a cost center with a pulse. That belief is either load-bearing in how you lead or it is decoration. This conversation is where you find out which.

Practical Steps

  1. Put the conversation on the calendar this week, before you feel ready. You will never feel ready. Waiting for clarity is the failure mode, not the safeguard. Thirty minutes, this week, and tell them the topic in advance so they can think instead of just react.

  2. Do a task-level inventory before you talk, not a role-level one. Take each role and break it into its actual weekly tasks. Mark each one: likely automated within twelve months, unlikely, genuinely uncertain. If you cannot fill this out, you are not ready to make headcount decisions either, which is worth knowing.

  3. Say the scary story out loud and name it accurately. Do not strawman it. “You’ve probably seen that AI was the top cited reason for U.S. layoffs four months running. That’s real. Here’s what’s true here.” Acknowledging their fear is what earns you the right to add context to it.

  4. Give one number and one date. Vague comfort evaporates in the parking lot. “No AI-driven headcount reductions through year end, and you hear it from me first if that changes” survives the drive home. If you cannot give a number, say why you cannot, and give a date by which you will.

  5. Ask what they have already decided. Straight question: “What have you concluded about your job that you haven’t said to me?” Then stop talking. The silence will be uncomfortable. Sit in it. What comes out is the actual state of your company.

  6. Fund the upskilling on company time, and be specific about it. Mercer found 63% of employees would trade a 10% pay raise for AI and digital upskilling opportunities. Read that twice. Your people are asking to be made more valuable. Hours on the calendar, not a link to a course library.

  7. Repeat it quarterly, even when nothing has changed. Especially when nothing has changed. “Still no change, still committed” is a message. A single brave conversation followed by eleven months of silence just resets the clock on the fear.

Frequently Asked Questions

What if I honestly don’t know whether AI will affect my team’s jobs?

Say that, precisely. “I don’t know” plus a task-level map plus a date to revisit is credible. “I don’t know” alone is not. Uncertainty communicated with structure reads as honesty; uncertainty communicated as silence reads as bad news you are sitting on. Your team can handle not knowing. They cannot handle being managed.

Won’t talking about AI and jobs create panic that wasn’t there?

The fear is already there. Mercer’s 2026 survey found 40% of employees worldwide worry AI will make their job obsolete, up from 28% in 2024, and 62% say leaders underestimate the emotional weight of it. You are not introducing the topic. You are joining a conversation that started without you, months ago.

Should I promise nobody will lose their job to AI?

Only if it is true and you can hold to it. A promise you break is worse than the silence it replaced, because it destroys the credibility you will need for every message after it. Promise the process instead: how decisions get made, how much notice people get, who tells them, and what support exists. Those are promises you can actually keep.

Are AI layoffs real or just an excuse for cuts companies wanted anyway?

Both, and the mix matters. Challenger, Gray & Christmas recorded AI cited in 101,743 U.S. job cuts through June 2026, roughly 23% of all cuts. But MIT’s Paul Osterman argues much of it is “AI washing,” a cover story for decisions driven by ordinary financial pressure. Gartner found firms cutting deepest saw no better returns than those cutting least.

Isn’t rolling out AI tools the more urgent priority right now?

Tools rolled out to an anxious team get quietly sabotaged, not adopted. People do not enthusiastically train their replacement. The conversation is not a delay before the rollout; it is the precondition for the rollout working at all. Mercer found only 44% of employees are thriving at work, down from 66% in 2024. Depleted people do not adopt anything.

The Close

I think about that conversation with my team member a lot. The nod. The moving on to the next agenda item. How efficient it felt.

I had told myself I was protecting her from uncertainty. I was protecting myself from a hard thirty minutes. Those look identical from the inside, and only one of them is leadership.

Here is what I want you to hear. The economists are going to keep arguing. Yale will publish, Wharton will publish, the headlines will swing from apocalypse to nothingburger and back, probably twice before Christmas. You are not going to out-forecast Wharton and you do not need to.

Your assignment is smaller and infinitely heavier. There is a person who works for you who lies awake wondering if they are next, and who has decided, based on your silence, that the answer is probably yes. They have not told you. They will not tell you. They are being polite. And every week you wait, that story hardens into something you will not be able to talk them out of later, no matter how good your eventual answer is.

You have the answer they need. Not the big answer. The small one. What is actually true about their actual work at your actual company. That answer costs you thirty minutes and some courage, and it is worth more to them than every tool you will deploy this year combined.

The scary story is winning by default. Default is a decision. Make a different one this week.

Say something. They are listening harder than you think.


About the author

Jonathan Mast is the founder of White Beard Strategies, where he provides AI coaching and mentorship to entrepreneurs who want to use these tools without losing what makes their business worth building. He is a speaker, a builder, and the creator of the Perfect Prompt Framework. He writes about AI, leadership, and the parts of both that nobody posts about.