For HR leaders asked to drive AI adoption and protect wellbeing at the same time, the message is pointed. The conditions doing the most damage are set by decisions inside the organization, and many of those decisions sit with People teams.
New data from The Well Minds Report 2026 shows that the biggest drivers of mental strain at work are decisions HR leaders already own: how AI is rolled out, how job security is communicated, and which benefits actually earn their place.
Nearly half of full-time employees (46%) are under mental strain, according to The Well Minds Report 2026, new research from Project Healthy Minds Workforce Lab and Wellhub. The single strongest predictor of that strain isn't pay, sleep or job security. It's pressure from leadership to use AI.
Among the one in ten employees who feel "a lot" of that pressure, 82% are under strain. The report analyzed June 2026 survey responses from 5,021 full-time employees in ten countries and ranked 70 characteristics of employees and workplaces by how strongly each one predicts strain.
For HR leaders asked to drive AI adoption and protect wellbeing at the same time, the message is pointed. The conditions doing the most damage are set by decisions inside the organization, and many of those decisions sit with People teams. Here are the three that matter most.
The Well Minds Report counts an employee as under strain when at least two of four things are true: they rate their mental health as poor or worse, they experience burnout at least weekly, they've felt emotionally checked out of work in the past year, or their work stress has risen over the past twelve months.
By that measure, 46% of full-time employees are under strain, according to The Well Minds Report 2026. It isn't concentrated at the front line. Strain runs from 42% of individual contributors to 55% of executives, which means the leaders you're asking to model healthy behavior are, on average, more strained than the teams they lead.
What makes the report useful for practitioners is how it ranks causes. Many risk factors travel together (employees who feel pressure to use AI are also more likely to be senior, remote and on shrinking teams), so the researchers used a model that holds all 70 characteristics constant at once. The "all else equal" figures below show how much more or less likely an otherwise identical employee is to be under strain.
The obvious suspicion is that AI itself is wearing people down. The data points somewhere more actionable.
Employees who feel a lot of pressure from leadership to use AI are 15 percentage points more likely to be under strain than otherwise identical colleagues who feel none, according to The Well Minds Report 2026. Nothing else in the survey, including money and job security, carries an effect that large. The same pattern holds for burnout and for feeling psychologically unsafe at work.
Two details make this especially relevant for whoever owns change management:
The report also offers clues about what turns the same expectation from a demand into an invitation. Employees who feel equipped to work with AI are far more likely to feel resilient and able to adapt (87% against 45%). And 77% of employees at companies with a clear, well-communicated AI policy feel resilient and able to adapt, compared with 39% of those unsure a policy even exists, according to the report.
The decisions you own here:
Only 7% of employees say their job security got worse this year, but those who do are 13 percentage points more likely to be under strain, all else equal, according to The Well Minds Report 2026. Fear of AI replacing one's job is far more widespread (38% of employees) and adds another 7 points.
Put that next to the AI finding and a pattern emerges. A meaningful share of AI pressure does its damage through job insecurity. When leadership pushes adoption without saying what it means for roles, employees fill the silence themselves.
The report names clearer communication about job security as one of the levers employers can pull. Most HR leaders already know this conversation matters. The harder questions are about execution: what you can say honestly when the roadmap isn't settled, how early to say it, and whether the message lands better from the CEO, a direct manager or People leadership.
Money compounds the picture. Nearly half the workforce (47%) says finances are stopping them from investing in their own wellbeing, and those employees are 14 points more likely to be under strain, the largest risk in the survey after AI pressure.
This is the finding most likely to spark debate in a total rewards meeting. Once everything else about an employee is held constant, the number of wellbeing programs an employer offers has no link to strain, according to The Well Minds Report 2026. Two programs or ten, the model can't tell them apart.
Only two programs sit on the protective end of the ladder:
What these two have in common is that each removes a practical barrier. The strongest protection in the entire study comes from the basics: physical activity at least weekly (−6 points, a larger effect than any manager behavior or culture feature measured), healthy eating and seven hours of sleep (−4 points each). Employees already know this. When asked what stops them exercising, they name time (52%) first. When asked what's hurting their mental health, insufficient sleep (39%) comes first.
That gives benefits leaders a useful test to run on the rest of the menu: does this program remove a real barrier of time or money, or does it just add a line to the benefits guide? And because time is the top barrier, the most effective wellbeing investment may not be a program at all. It may be protected hours, limits on after-hours work and a culture where leaving to work out isn't read as slacking.
The report is candid about its limits, which is part of what makes it worth taking seriously. The survey is a snapshot, so it shows what goes together rather than proving cause and effect. The AI items measure how pressure and policy are perceived, not what leaders actually did. Respondents came from an online panel with equal country quotas and no weighting. The researchers stress-tested the ranking five ways, including a machine-learning model that reproduced the same ordering, and the AI pressure effect held throughout. For practitioners, that's the right level of confidence: strong enough to change how you sequence an AI rollout, not a reason to skip asking your own people.
The report is clear on what drives strain and deliberately open on what to do about it. Whether the right response to AI pressure is training, clearer communication about job security or a different messenger, it says, depends on the organization. The same goes for which benefits to cut, keep or expand.
That's exactly the kind of judgment call where a peer who has already been through it is worth more than another framework. If you're a Troop member, bring these questions to the Peer Community Forum:
The report's closing advice applies to all of it: support that leaders visibly practice reads as an offer, and support they skip reads as a demand.
________________________________
What is The Well Minds Report 2026? The Well Minds Report 2026 is a research report from Project Healthy Minds Workforce Lab and Wellhub that examines what drives mental strain at work and what employers can do about it. It analyzes survey responses from 5,021 full-time employees in the United States, United Kingdom, Brazil, Argentina, Chile, Spain, Italy, Germany, Mexico and Romania, collected in June 2026 for Wellhub's Work Life Wellness 2027 study.
What does "under strain" mean in the report? An employee is counted as under strain when at least two of four criteria apply: poor or worse self-rated mental health, weekly or more frequent burnout, feeling emotionally checked out of work in the past year, or rising work stress over the past twelve months. By this measure, 46% of full-time employees are under strain.
What is the strongest predictor of employee mental strain? Perceived pressure from leadership to use AI. Employees who feel a lot of pressure are 15 percentage points more likely to be under strain than otherwise identical colleagues who feel none, the largest effect of the 70 characteristics tested.
Does offering more wellbeing benefits reduce employee strain? Not by itself. The report finds the number of wellbeing programs has no link to strain once other factors are held constant. Only fitness programs (−4 points) and financial wellness tools (−5 points) register as protective, and both work by removing practical barriers of time or money.
How can employers reduce strain from AI adoption? The report points to how the expectation is set, what leadership says about job security and how much support employees get to learn. A clear, well-communicated AI policy is linked to resilience: 77% of employees at companies with one feel able to adapt, compared with 39% of those unsure a policy exists.
.jpg){kind=spacer}
.jpg){kind=spacer}
.jpg){kind=spacer}
.jpg){kind=spacer}
.png){kind=spacer}
.jpg){kind=spacer}
