Anxiety at work isn’t a vibe. It’s a trend line, and it’s still moving in the wrong direction. In 2026, 48% of US adults said they feel more anxious than they did a year earlier — up from 43% in 2025 (American Psychiatric Association). Workplace stress backs that up: 72% of US workers now report moderate-to-high stress, the highest figure in seven years, and 55% describe themselves as burned out, a six-year high (Aflac; Eagle Hill Consulting).
This didn’t start with the pandemic, and it isn’t a cyclical blip. Jean Twenge’s generational research — drawing on national surveys stretching back to the 1930s — found anxiety and depression rates among young people held roughly steady for decades, then broke sharply upward starting around 2011–2012, tracking the mainstream adoption of smartphones and social media. Economic anxiety has compounded it: 75% of US adults now cite the economy as a significant source of stress, the highest level recorded in 17 years.
What’s actually driving it
Three forces are compounding rather than operating independently. Economic uncertainty is directly associated with higher rates of anxiety and depressive disorders across nearly every adult age group. Social media adds a layer of constant social evaluation and information overload that depletes the cognitive bandwidth people would otherwise use to regulate stress. And societal division has become its own stressor: 62% of US adults cite it as significant, which is why the APA titled its 2025 national survey “A Crisis of Connection.”
The layer HR leaders can’t ignore: AI
AI-driven job uncertainty is now a measurable, independent driver of workplace anxiety — not a background worry. 40% of employees fear losing their job to AI, up sharply from 28% just two years ago (Mercer Global Talent Trends). 69% believe AI will lead to layoffs at their company within three years, and globally, only 22% of workers strongly agree their job is safe from elimination (ADP Research). Nearly a quarter say AI is actively harming their mental health today.
The most telling data point for HR functions specifically: 58% of workers say they’d feel safer discussing their mental health with a chatbot than with their own HR department — up from 50% just a year prior. That’s not an AI problem. It’s a trust problem, and it’s getting worse.
The buffer: connection, not just coping
Here’s where the research gets actionable. Loneliness doesn’t just correlate with anxiety — longitudinal studies show it predicts elevated anxiety up to 21 years later. The APA’s 2025 data makes the size of the effect concrete: 80% of adults with high loneliness report living with chronic illness, including anxiety disorders, compared to 66-68% among those with low-to-moderate loneliness.
The mechanism has been understood since Cohen and Wills’ foundational 1985 “buffering hypothesis”: social support doesn’t just make people feel better in the abstract — it measurably weakens the link between a stressor and its negative health impact. Harvard’s Study of Adult Development, now in its 88th year, found that relationship quality at age 50 predicted long-term physical health better than cholesterol levels did. Social fitness outperformed genetics, income, and IQ as a predictor of a long, healthy life.
This holds inside the workplace too, with numbers HR teams can act on directly. Gallup’s research finds employees who strongly agree they have a best friend at work are seven times more likely to be fully engaged. Organizations that move the needle on workplace friendship — getting six in ten employees to report a close work relationship — see measurable downstream effects: a 28% drop in safety incidents and meaningfully higher profitability.
The implication
As AI uncertainty and economic pressure keep pushing anxiety upward, the instinct is often to respond with more communication, more reassurance, more policy. The research suggests the more durable lever is structural: investing in the quality of manager relationships, peer bonds, and psychological safety isn’t a soft benefit sitting alongside performance — it’s one of the few interventions with a documented, measurable buffering effect on the anxiety this moment is producing.
Sources
- APA: Annual Mental Health Poll 2026
- APA: Stress in America 2025 — A Crisis of Connection
- Aflac 2025 WorkForces Report — via SpringWorks
- Eagle Hill Consulting burnout survey — via Metaintro
- Jean Twenge research overview — The Anxious Generation
- Economic uncertainty and mental health, 1991–2019 — NIH/PMC
- Mercer Global Talent Trends 2026 — via Metaintro
- ADP Research Institute — via Fair Play Talks
- Loneliness and anxiety, longitudinal evidence — ScienceDirect
- Cohen & Wills (1985), Stress, Social Support, and the Buffering Hypothesis
- Harvard Study of Adult Development — Harvard Gazette
- Gallup: Best Friend at Work
Want to know where your organisation stands? Get our free Employee Connection & Productivity Audit — a 10-question tool for HR leaders.