The Geography of AI Anxiety

Do we need AI regulation, or something else like labor market reform?

Theres an awful lot of anxiety out there about AI. Will it take my job or make the skills I’ve developed over twenty years redundant? Will it cut my hours or turn me into a drone?

The fact is, major technological disruptions always generate a lot of angst among us worker bees. It’s the same with AI… only more so. The incredible speed of AI development and adoption, combined with it’s power makes it particularly unsettling for many people. I’d be lying if I said I didn’t feel it (at least a little bit) myself.

But should we really be worried? If so, how worried should we be? My view is, we need to look at what the data says before running through the streets with our hair on fire.

TL/DR: Possibly. But probably not as much as the headlines would lead us to believe. The fact is, worker anxiety about AI displacement varies (often dramatically) by country but the patterns we see are both easily discernable to some degree predictable. The truth is, anxiety about AI generated job losses seems to track pretty closely with specific structural features in each labor market.

So what’s driving the fear in the US? And how different – or similar – are other developed markets in their response to this transformative technology? Finally, while anxiety levels have historically declined as people became more familiar with new technologies –this isn’t happening as quickly with AI. In the US, at least, it seems to be increasing.

My initial thought is that the different attitudes towards AI and the threat of AI-driven job losses is probably driven more by labor market structure and peculiarities than anything else. There are other factors for certain, but the main driver (at least on the US) is the relative lack of employee protections.

However, let’s look at the data we have before drawing any firm conclusions.

The US Data

American workers seem to be feeling AI job anxiety more acutely than most other developed countries. Not only that, the feeling of dread seems to be increasing over time, not decreasing:[1] [2] [3]

89% of US workers surveyed in January 2025 reported AI-related concern about job security
43% say they know someone who has already lost a job because of AI
60% of American workers believe AI will eliminate more jobs than it creates
Only 22% say leadership has explained how AI will be applied to their work (Gallup)
Mercer survey of 12,000 workers: 40% fear losing their job to AI, up from 28% a year earlier

The UK: The Exception that Proves the Rule

The UK is the natural test case for the “labor market structure drives anxiety” thesis, and the data seem to support it.

Over 70% of UK respondents voiced concern about AI-induced unemployment (UK Tracker Survey 2024) — far closer to the US than to continental Europe[4]
UK labor market is structurally more similar to the US than the EU: weaker employment protections, no continental just-cause requirements, more flexible hire/fire environment
Post-Brexit, the UK is not subject to EU AI Act protections

To test this assumption – that the structure of the labor market is a key driver for AI anxiety about employment – let’s ask: What if it’s not structural, but is mainly a matter of culture or familiarity with the technology? If that was the case, the UK should track closer to France or Germany, but it doesn’t by a wide margin. If anything, it tracks closer to the US.

If we accept Occam’s Razor as a good way to analyze competing ideas (the simplest explanation is usually the right one), the labor market structure thesis remains the best explanation for why the UK and the US seem to have more similarities than differences.

The EU

Of all the developed markets, countries in the EU show the sharpest contrast with the US and the UK. The GlobeScan Gen Z survey[5] revealed startlingly low levels of concern about the dangers of AI. The numbers below represent respondents who said that the dangers of AI (including for job losses) was “Very Serious”:

Germany: 24%
Spain: 20%
France: 18%
Sweden: 18%
Netherlands: 16%

It’s worth noting that the GlobeScan data specifically covers Gen Z respondents; comparable all-age EU surveys show similarly low figures, suggesting the pattern holds across demographic groups. Shared labor market standards and structures that differ significantly from the US[6] seem to explain this relative lack of serious concern on the part of most EU countries:

Just-cause termination is the legal standard across continental Europe, and AI adoption alone cannot constitute grounds for dismissal the way it can in the US or even the UK;
The EU AI Act imposes transparency requirements on AI-assisted workplace decisions[7], reducing the “something is happening and no one told me” dynamic;
CEPR research[8] shows that AI adoption has actually increased EU labor productivity around 4% with no evidence of reduced employment in the short run; and
Broader social safety nets (minimum four weeks paid leave, stronger unemployment benefits, retraining programs) reduce the downside risk of displacement

China: Anxiety Compounded by Structural Fragility

China offers the most complicated (and opaque) narrative on this issue. Claims that Chinese workers aren’t anxious about AI appears to be wrong, or at best outdated – though the GlobalScan Gen Z survey shows that at least some segments of China’s workforce don’t view AI as a serious threat. There is also the fact that much of China’s internally generated data can (and should) be viewed with a healthy degree of skepticism. .

Youth unemployment (16–24, not in school) hit 18.9% officially in August 2025;
Job postings for college graduates fell 22% in the first half of 2025 vs. the year before (this is from major Chinese recruitment site data);
In 2024, the city of Wuhan saw protests over Baidu robotaxis displacing drivers (public protests always capture significant public and national attention in China); and
A DeepSeek representative warned publicly that AI disruption could trigger a labor crisis and “shake society to its core” (public warnings like this from private companies is highly unusual in China; it could not have been issued without government approval).

One thing we can say is that AI anxiety in China is real[9], but it is almost impossible to isolate from other pre-existing structural crises. Some of those ongoing challenges are a national collapse in real estate values, a contraction of the tech sector, and the “lying flat” phenomenon among young workers.

Finally, the ruling Chinese Communist Party’s (the CCP) “AI+” plan targets 70% AI penetration by 2027 and 90% by 2030. The pace of acceleration could easily exacerbate the problem. Stayed tuned – if this happens, it could be global (not just Chinese national) news.

Remember: TheCCP has structural reasons to suppress or manage public discourse on job losses, so reported sentiment data may undercount real anxiety. However, signs are emerging that Beijing is trying to scale down AI adoption rates to avoid social unrest. This includes delivering an “AI + Employment” framework that includes tax incentives, wage subsidies, and limits on AI displacement of certain jobs. Clearly, the CCP is aware of the challenges even if everyday citizens aren’t able to talk openly about them.[10]

Japan: The Paradox of Low Adoption and High Anxiety

Japan presents a counterintuitive case that challenges my admittedly simple framework of labor market structures as the key driver for AI anxiety:

Only 8.4% of Japanese workers currently use AI tools at work[11]; this is among the lowest adoption rates in any developed economy;
Only 23.5% of Japanese SMEs report using generative AI (vs. 38.7% in Germany)
And yet: only 5% of Japanese workers feel their jobs are secure. the lowest figure of any surveyed country (ADP Research, 2026)[12]
73.8% of Japanese workers in finance/insurance worry about job loss over 10 years vs. 68% globally

My interpretation is still on the structure of the labor market, but with a peculiar Japanese twist: Japan’s traditional lifetime employment system created strong job security expectations is eroding just as AI arrives.[13] Attitudes are changing among younger workers; currently, only 30.1% of new Japanese employees now aspire to stay with one company long-term. This used to be a near universal expectation just a generation ago. This is a significant psychological blow for workers who expected a specific form of security (long-term tenure with a single employer) but who are beginning to lose it. It doesn’t help that AI is arriving at the same time. We all know that correlation is not causation – but the proximity certainly doesn’t help matters.

On top of all this, Japan is facing a genuinely threatening demographic inversion. The working age population is shrinking rapidly, and Japanese society is experiencing acute labor shortages across every industry including caregiving, services, and manufacturing. AI has been identified as both a threat to some workers and a structural necessity to fill gaps that humans can no longer fill.

South Korea: High Anxiety, Staying Competitive

South Korea is the most interesting outlier in the countries we looked at. There, anxiety is high, but its character is significantly different from what we see in the US.

78% of South Korean office workers using AI fear falling behind (vs. 65% global average, per Microsoft 2026 Work Trend Index)[14]
64% fear AI could displace human labor and exacerbate inequality
While the Korea Development Institute estimates AI may replace roughly 256,000 workers per year over the next decade, this number is well below the projected annual decline in Korea’s working age population (about 390,000 annually over the same period).

At the end of the day, Korean anxiety appears to be more about skills obsolescence and falling behind peers than about immediate job displacement. The fear is competitive rather than existential — “will I be able to keep up?” rather than “will I still have a job?” This appears to be related to Korea’s intensely competitive education and credentialing culture.

As a counterpoint, the MIT Technology Review (June 2026)[15] found that South Koreans actually “love AI” at a higher rate than peer nations. This is not a population hostile to the technology; they’re anxious specifically about being left behind in a race they actually want to run. The issue of inequality is also one that, at least culturally, may drive more opposition than most survey’s can reasonably project.

The Three Main Contenders

The world is never binary, and it’s always ambiguous (at least to some degree). While we still believe that the primary driver of AI-driven job loss anxiety is the structure of labor markets, there are other important factors we need to consider as well. These span culture, politics, demographics, and other variables. But three variables seem to dominate and offer the simplest explanation:

1. Labor market structure (strongest predictor)

This is still the strongest predictor of how much anxiety workers will experience across countries. Geographies with strong just-cause dismissal requirements (continental EU) show the lowest anxiety, while the US and the UK (with far weaker worker protections) feel the highest levels. This isn’t a coincidence but a rational risk assessment by workers.

2. The communication gap

The US Gallup data (44% of employees know AI is being used; only 22% have been told what it means for their jobs) may be uniquely bad by global standards. The EU AI Act’s transparency requirements are likely narrowing this gap in Europe. This is a hypothesis worth flagging where data gaps exist for Japan and Korea.

3. Anxiety type matters as much as anxiety level

Not all anxiety is equal, and it doesn’t always come from the same source. The character of AI anxiety differs by country:

US/UK anxiety: about job loss
Korean anxiety: about falling behind
Japanese anxiety: about a social contract collapse of which AI is one component
Chinese anxiety: entangled with pre-existing structural economic failure

All tigers are cats, but not all cats are tigers. In other words, we see anxiety about AI’s impact on work and employment almost everywhere, but the sources of anxiety and worry are rarely identical.

Conclusion

First: High US AI anxiety isn’t irrational or technophobic, it is structurally predictable given a labor market where workers have the fewest protections against displacement of any developed world economy. But the anxiety is mainly a signal; the real question is, what should we do with it? I don’t have any answers; if I did, I’d be a rich man. What I can suggest, however, is the policy makers and decision makers inside companies owe their employees honesty and straightforward answers. It all starts with an honest and data driven conversation around the issue – not political posturing. (Fat chance, may of you will say, but we can always hope…?)

Second: Countries with the lowest anxiety scores aren’t less exposed to AI. Nobody in today’s advanced economies is, but they are better insulated from its worst potential consequences. Here, the difference isn’t the technology, it’s the social architecture that has been built around work. That architecture has become embedded in the culture and informs what people assume about the future.

Third: Domestic politics, history, and culture directly shape attitudes towards – and anxiety about – AI and its impact on job security and employment. Remember what we see in Japan, Korea and China:

In Japan, very few workers feel secure as AI is adopted more broadly – even though it is far from being the main driver for changes to the labor market and job security. It’s become an easy target to which people attach their insecurities.
In Korea, the issue is less about job security and more about ‘keeping up’ and staying competitive.
In China, it’s incredibly difficult to accurately gauge what people really think – but we can tell there is significant anxiety in the way the CCP is ‘softening’ it’s official policies about AI implementation.

Finally, here’s what is potentially my most controversial suggestion: if AI anxiety in the US actually does track tightly with the relative lack of labor market protections, perhaps the most effective policy response to AI displacement may need to look less like AI regulation and more like labor market reform. I’d love to hear people’s thoughts on this.

[1] Resume Now AI Disruption Report (Jan 2025)

[2] GrindHotline — AI Job Loss Fear Statistics 2026

[3] JFF Survey (2025) — Worker Anxiety Over AI Is Growing

[4] UK 2024 AI Tracker Survey data (via GrindHotline)

[5] GlobeScan — Gen Z AI Risk Perceptions (May 2026)

[6] RTE Brainstorm — Why the US and Europe differ on job security

[7] OnLabor — Europe Is Regulating AI Hiring. Why Isn’t America?

[8] CEPR — How AI is affecting productivity and jobs in Europe

[9] RAND — China Is Worried About AI Job Losses (Dec 2025)

[10] Matt Sheehan Substack — China is getting worried about AI & jobs

[11] OECD — AI Use in the Japanese Workplace

[12] ADP Research Today at Work 2026 (Japan data)