AI anxiety at work: what leaders can do

What is AI anxiety?

AI anxiety is worry about what AI means for your job, your skills and your place at work. It usually comes down to three questions. Will I still have a job? Will I keep up? Will I still matter? It’s an everyday term, not a medical diagnosis.

Outside work, the fear shows up in public polls too. In June 2025, 50 percent of US adults told Pew Research Center they were more concerned than excited about the growing use of AI in daily life, up from 37 percent in 2021.

Is it rational to fear AI at work?

Partly, yes. Some tasks will disappear and some roles will change, and nobody can promise otherwise. That’s why reassurance on its own falls flat. People notice when they’re being calmed down instead of informed. If you work in HR and wonder about your own role, read Will AI take HR jobs?

How common is AI anxiety at work?

Job loss is only part of it. In Pew’s survey, 52 percent of workers were worried, but 32 percent expected fewer job opportunities for themselves (5,273 US workers, October 2024). An answer that only covers job security misses much of the worry.

Fear of job loss is higher where AI is already in use. Gallup’s 18 percent rises to 23 percent in organizations that have adopted AI (23,717 US employees, February 2026). In Great Britain, 26 percent of employees worried that AI will lead to job losses (YouGov for Acas, 1,023 employees, spring 2025).

The worry also tends to come with strain. In a 2023 survey of 2,515 US workers for the American Psychological Association, those who feared AI might make some of their duties obsolete were more likely to feel stressed during the workday, 64 against 38 percent. It’s a correlation, but reason enough to take the worry seriously.

What drives AI anxiety?

The worries fall into seven groups, and each needs a different answer.

The training gap is wide. In BCG’s 2026 survey of 11,749 people in 14 markets, 88 percent expected to need major upskilling within five years, but only 36 percent felt properly trained. A course alone won’t close that gap. People need time on their own tasks and a colleague to ask, which is what AI enablement with a champion in each team sets up.

On monitoring, the AI Act already draws one clear line. Since 2 February 2025, it bans AI systems that infer the emotions of people at work, except for medical or safety reasons (Article 5).

What should leaders say? A conversation guide

Your own behavior as a manager counts. In Gallup’s 2026 data, employees whose manager actively supports the team’s AI use had an engagement rate of 48 percent, against 30 percent among the rest. That’s an association rather than proof, and I’d still act on it.

It’s tempting to wait until you have all the answers. Don’t. People fill silence with rumors, and the rumors are usually worse than the plan. Have the first conversation with what you know, in a regular team meeting of about 45 minutes.

1. Prepare three lists (before the meeting, 30 minutes). Write down what’s decided, what isn’t decided yet and what nobody knows yet. Check with HR and your own manager what you can say about roles and jobs.
2. Open with the plan (5 minutes). Say why the organization uses AI, what changes first and what stays the same for now. Skip the hype.
3. Ask before you answer (10 minutes). Ask “What’s your biggest question about AI in our work?” If the team is quiet, hand out notes and let people write their question without a name.
4. Sort the questions into three piles (15 minutes). Use the piles from your lists, “decided”, “not decided yet” and “nobody knows yet”. For the not-decided pile, give a date for when you’ll come back.
5. Agree on one small step (10 minutes). Pick one task the team tries together in the approved AI tool within two weeks.
6. Follow up in writing (within a week). Send the questions, your answers and the open items, and keep the dates you gave. Offer a one-to-one to anyone who’d rather talk alone.

Talking points for managers

Adapt these to your team, and only use the lines you can stand behind.

Here's what's decided. We're starting to use [tool] for [tasks], from [month].

Here's what isn't decided yet. [For example, how we'll handle X.] I'll come back to you on that by [date].

Here's what nobody knows yet. That includes exactly what every role looks like in three years. If decisions are made that affect your role, you'll hear it as early as I'm allowed to share it, and you'll hear it from me.

Here's what we won't do. [For example, use AI to assess individual performance.]

Here's what I'm asking of you. Try it on one real task in the next two weeks and tell me what worked and what didn't. You have [time, I'd start with 30 minutes a week] for that during working hours.

And here's what I'll do. I'll use it myself, share my mistakes and take your questions to [leadership team or HR].

Answers to the hardest questions

What should leaders not promise?

A promise you can’t keep costs more trust than an honest “I don’t know.” These are the tempting ones.

A no-layoffs pledge is a decision for the board or the executive team, and a manager shouldn’t improvise one. If your organization has made one, say it plainly and say for how long. If it hasn’t, don’t hint at one. For what the executive team needs to decide first, see jobs and headcount in an AI transformation.

How do you involve people in AI decisions?

I think people push back much harder on an AI rollout that’s done to them than on one they helped shape. So give them a real say in where AI goes first.

• Start with the work. Let each team map its recurring tasks and choose the first ones to try, with the task sort below.
• Let volunteers go first. Give the curious ones time and a chance to show colleagues what worked. My guide to AI for leaders covers how to back them.
• Bring employee representatives in early. The AI Act adds a duty to inform for some AI. From 2 December 2027, employers that put a high-risk AI system into use at work, such as AI for recruitment or for decisions about workers, must first inform workers’ representatives and the affected workers (Article 26(7)). National law or collective agreements may already require you to inform, consult or negotiate earlier.
• Put leaders in the room. When my colleagues at humAIn Academy and I trained about 80 employees at Mediemyndigheten, the Swedish Agency for the Media, the leadership team joined the employee groups instead of taking a separate course. They heard the concerns first-hand and used them to shape strategy and policy.

Team exercises that lower AI anxiety

Talking helps, and trying things together helps more. A manager can run these three without outside help.

The task sort (45 minutes)

Each person lists ten tasks they do every week or month and sorts them into three columns, “Keep doing it myself”, “Let AI help me” and “Not sure”. Compare as a team and look hardest at the tasks where people disagree, because that’s where the worries sit. End by choosing two tasks to try first.

Try it together (60 minutes)

Pick one low-risk task the whole team does, like summarizing meeting notes or drafting a standard reply. Everyone does it in the approved tool, then you compare results and look for what the AI got wrong. I think seeing it fail on a familiar task builds more confidence than a polished demo.

The time deal (20 minutes)

Agree where saved time goes. Some to the work you never get to, some to learning, some to breathing room. Write it down and look at it again after a month. In BCG’s 2026 survey, two in three regular AI users without a management role (BCG calls them frontline employees) got limited or no guidance on what to do with the time AI saves.

Which signals should leaders watch?

AI anxiety rarely shows up as a complaint. Watch for these.

• Silence. If a team that AI affects never asks about it, assume the questions are going somewhere else.
• Hidden use. People use AI without mentioning it, or in personal accounts. See my guide to shadow AI.
• Paid licenses, little use. It can mean people are unsure what’s allowed or afraid of looking replaceable.
• Unanswered rumors. Talk of monitoring or cuts that nobody in leadership has addressed.
• Longer days. People save time with AI and still work more.
• Your strongest people asking where their role is going. Talk to them before they go looking for answers elsewhere.

Three questions for a quarterly pulse

1. I know what AI means for my work over the next year.
2. I have the time and support to learn AI on my own tasks.
3. I can raise concerns about AI without it counting against me.

Use a 1 to 5 scale, add one open question (“What would make AI more useful, or less worrying, for you?”) and follow the trend team by team, for example in your engagement survey. If the third question scores low, I’d act on that first. For a shared way to talk about where each person is in the change, see the ADKAR model and the change curve in an AI rollout.

What if the anxiety is your own?

Managers and HR people feel this too. Here’s what I’d do.

• Name the specific worry. “I’ll lose my job” and “I can’t keep up” need different answers. If it’s the first, check how exposed your own job is, task by task.
• Ask your manager three questions. What does AI mean for my work this year? What’s expected of me? How much time do I get to learn?
• Try it on something small. Use the approved tool on one low-risk task a week and note what it got right and wrong.
• Pick your sources. One newsletter or colleague you trust beats a feed of headlines.
• Get help if it lasts. If the worry affects your sleep or health over time, talk to a doctor, occupational health or your employee assistance program.

What should HR do about AI anxiety?

Three things I’d have HR do first.

• Equip managers before the rollout with the talking points and answers above, and train them on the tools. AI training for managers is where I’d start.
• Write down what AI is used for in decisions about people, such as recruitment, performance and scheduling, and share the list with employees. The template below is a starting point.
• Treat a large AI rollout as change in your stress risk assessment. The UK’s HSE counts how change is managed and communicated among six areas of work design linked to poor health when badly managed. Tell managers where to send someone whose worry turns into sleep problems or sick leave, such as occupational health or the employee assistance program.

Where we use AI in decisions about people (template)

Fill it in with HR, share it with the team and update it before anything changes.

End the list with “Questions go to [name]. Last updated [date].” If a row covers recruitment or decisions about workers, check whether the system counts as high-risk under the AI Act, which brings the duty to inform described above from 2 December 2027.

What Swedish HR respondents report

Among HR respondents in Sweden, active involvement of employees is still uncommon. In the AI & HR Barometern 2026, humAIn Academy’s survey of 329 HR professionals in Sweden who chose to respond between 8 June and 24 September 2026 (likely more interested in AI than HR overall), 16 percent said employees or union representatives had been actively involved in questions about AI use at work, and another 32 percent said they had been informed. Lack of skills was named among the biggest obstacles to using more AI in HR by 40 percent, resistance or culture by 17 percent. So give people time to learn on their own tasks before you read hesitation as pushback.

Frequently asked questions

What is AI anxiety?

AI anxiety is worry about what AI means for your job, your skills and your place at work. It’s an everyday term, not a medical diagnosis. It usually comes down to three questions. Will I still have a job? Will I keep up? Will I still matter?

How do you talk to employees who are afraid of AI?

Say what’s decided, what isn’t decided yet and what nobody knows yet, with a date for the open items. Ask for their questions before you answer, agree on one small task to try together and follow up in writing within a week.

Should leaders promise that nobody will lose their job to AI?

Only if the organization has decided it and can keep the promise. Otherwise, promise what you control, such as early information, a say in how AI is used and time to learn.

What is an AI culture?

An AI culture is a workplace where people use AI openly, share what works and what doesn’t, know the rules and trust that the gains are shared. You see it when people ask questions instead of staying silent and managers use AI themselves. When it’s missing, you see the signals above, such as silence, hidden use and paid licenses nobody uses.

What can I do about my own AI anxiety?

Name the specific worry, ask your manager what AI means for your work this year and how much time you get to learn, and try the approved tool on one low-risk task a week. If the worry affects your sleep or health over time, talk to a doctor, occupational health or your employee assistance program.

Sources

• Pew Research Center: U.S. workers are more worried than hopeful about future AI use (2025)
• Pew Research Center: How Americans view AI (2025)
• Gallup: Rising AI adoption spurs workforce changes (April 2026)
• Gallup: AI in the workplace indicator
• Gallup: Employee engagement remains flat as AI adoption accelerates (July 2026)
• Microsoft and LinkedIn: 2024 Work Trend Index
• BCG: AI at Work 2026 (June 2026) and slideshow (PDF)
• American Psychological Association: Worries about AI at work (2023)
• Acas: 1 in 4 workers worry that AI will lead to job losses (2025)
• HSE: What are the Management Standards?
• AI Act Service Desk: Article 5, Article 26 and Article 113 (European Commission)
• State of AI in HR (AI & HR Barometern 2026, humAIn Academy)