Economic Outlook for Q2 2026

Quarterly Economics Briefing–Q2 2026

By Stephen Cooper, Patrick Coate, Yariv Fadlon, and Mari Villalta

Posted Date: June 25, 2026

Key Themes and Takeaways

  • It is still too early to draw conclusions on the impact of AI on employment
  • Employment trends in early 2026 appear to be negatively correlated with AI adoption rates, while productivity growth has been elevated
  • Higher productivity growth has helped support economic growth and wage growth


Artificial Intelligence (AI) and the Labor Market

“What impact will AI have on jobs?” We hear this question often among workers compensation stakeholders—and while it may seem straightforward, it’s anything but.

As Ethan Mollick, Wharton School professor and one of the most closely followed voices on AI, recently said:

“Nobody knows anything. I spend my time talking to AI labs, famous people, I talk to CEOs all the time, and nobody knows anything. We’re all making this up as we go along. So anyone who’s like, ‘We have the playbook’—they’re lying to you.”1

We’re not going to lie to you; we don’t know the exact answer either. We’re economic analysts, not clairvoyants, but we can look at past and present trends to help us prepare for the future.

Currently, economists hold a wide range of opinions on the future of AI. Some believe it will be a great enabler, ushering in a new employment boom. Others advise caution, believing this technology will cause significant displacement and eliminate jobs.

In this brief, we’ll explain why we believe the answer lies somewhere in the middle. We will explore how technology impacts jobs, what AI adoption has looked like so far, and what it could mean for the economy and workers compensation going forward.

Are You a Horse, a Travel Agent, or a Truck Driver?

You may have heard this statistic before: 80% of the U.S. workforce could have at least 10% of their work tasks affected by the introduction of LLMs (large language models) and, approximately 19% of workers may see at least 50% of their tasks impacted.2.

It’s one of the most widely cited statistics around AI’s impact on work. Many take it to mean that 80% of jobs are at risk. But that’s incorrect. When assessing the impact of AI (or any new technology) on employment, we want to focus on tasks, not jobs as a whole.

At its core, a job is a bundle of tasks. Some jobs are complex, with many discrete or interconnected tasks, while others have fewer, more straightforward tasks. This is, as economists call it, the dimensionality of a job.

You can find examples of complex bundles of tasks relatively easily—just think of registered nurses, project managers, and consultants, to name a few. These jobs will likely be exposed in some respects to AI due to the diverse number of tasks within each job bundle, but that does not mean these jobs are at risk of elimination.

Let’s say, for example, that a job has 10 unique tasks and AI can automate one of them (10% exposure, matching the cited research above). That frees up the worker to spend more time on the other 9 tasks, increasing the quality of outcomes, or to add a new 10th task. Both are examples of increased productivity, which can also result in higher wages because that worker is now more valuable to the company.

On the other side of the spectrum, some jobs have a smaller set of tasks or even just one task. These roles may be more vulnerable to new technologies. For example, many years ago, a horse performed a small number of very simple tasks: moving a plow across a field or moving a cart full of people or goods across town. Because of the simplicity of these tasks, the horse was easily replaced by machines (tractors and cars) History is full of examples where low-dimensionality jobs went the way of the horse. The industrial revolution saw the elimination of artisan weavers, ice cutters, and log drivers. Electricity saw the elimination of lamplighters. Early computing and mechanization saw the elimination of switchboard operators, elevator operators, and pin boys (yes, resetting the pins in a bowling alley used to be a manual job!).

But jobs with low dimensionality don’t necessarily have to go the way of the horse.

Travel agents almost went that way with the advent of the Internet. Some, however, survived by evolving. They did not fight the technology that was replacing them but instead shifted their tasks from booking trips (which was automated) to customer service, consulting, and problem-solving. Automated booking systems, after all, were not able to give experiential recommendations or solve last-minute travel emergencies. Some readers may see this as a silly example and think that travel agents are already obsolete. The shift from booking services to concierge services has indeed reduced the number of travel agents today compared to decades ago; however, this niche industry still employs over 70,000 workers and has grown year to date in 2026.

With that in mind, let’s consider a modern-day example of a job with relatively few tasks that is not just exposed to new technologies, but also at risk of elimination. Truck drivers, particularly long-haul ones, are a clear case. The main task of transporting cargo from point A to point B may soon be taken over by autonomous vehicles.

Does this mean that truck drivers are going the way of the horse? Maybe, but also maybe not.

If autonomous vehicles eliminate the task of driving between points, the truck driver may adapt and instead focus on other tasks or take on new ones. Drivers may become logistics planners; they may also take on vehicle maintenance or other job-related tasks, resulting in an evolution of the role rather than an elimination. When assessing the impact of AI—or any new technology—on work, a job’s dimensionality and the exposure of its tasks to the new technology matter. What also matters is whether automation allows the worker to focus on new or other tasks.

In fact, Jevons’ Paradox predicts that jobs with high dimensionality that are more likely to experience productivity gains because of AI may actually lead to increased employment. Jobs with low dimensionality, on the other hand, are more at risk of elimination and may require evolution on the part of the worker in order to avoid displacement.

The horse was eliminated, the travel agent evolved, and we will have to wait and see what happens to the truck driver.

Now that we have a conceptual framework for thinking about the impact of AI on employment, let’s shift to what we see in the data.

Uneven AI Adoption Across Industries

To explore the impact that AI may be having on the employment picture, let’s start with the U.S. Census Bureau’s Business Trends and Outlook Survey. This survey asks firms whether they have used AI in any business function in the previous two weeks, including applications such as machine learning, natural language processing, and virtual agents.

AI adoption rates vary across sectors, with the highest rates appearing in Management, Information, and in Professional, Scientific, and Technical Services. The lowest rates appear in Transportation and Warehousing and in Accommodation and Food Services. So how do these adoption rates impact employment? While AI-related layoffs started making headlines in 2025, we believe employment changes so far in 2026 better isolate AI’s influence. Employment in 2025 was heavily influenced by factors outside of AI—trade policy, immigration policy, and economic uncertainty, to name a few. The trend in employment growth emerging in 2026 appears to have shifted meaningfully from that of 2025. Through the first five months of the year, job growth averaged 114,000 jobs per month, compared to just 10,000 per month in 2025. Now, as the labor market recovers from a period shadowed by other factors, AI’s impact on employment trends may be easier to decipher.

First, notice that the early 2026 recovery has been uneven across sectors. Construction, Health Care and Social Assistance, Transportation and Warehousing, and Leisure and Hospitality have seen solid job growth, while office-based sectors such as Information and Financial Activities have continued to shed jobs.

Next, we compare employment growth by sector to reported AI adoption rates on the survey. While there are likely some lingering impacts from economic uncertainty, a clearer negative correlation has emerged in 2026 between sectors with higher rates of AI adoption and changes in employment.

With the exception of several outliers, job growth in early 2026 appears to be negatively correlated with AI adoption. Sectors reporting the lowest levels of AI adoption have experienced stronger employment growth, while sectors reporting higher levels of AI adoption have seen declining employment. This might not seem surprising at first glance, but it does give us an indication that in certain sectors, AI adoption may be contributing to employment declines.

The two outlier sectors, Health Care and Social Assistance and Professional, Scientific, and Technical Services, have higher levels of reported AI adoption but stronger employment growth than the trend relationship would suggest. These are both sectors with high-dimensionality jobs that may be benefiting from higher productivity, which in turn may be encouraging more hiring.

So, if AI adoption and hiring continue to reflect this early 2026 trend, the question for workers compensation becomes: How will this trend impact premium?

Employment growth in low AI exposure sectors such as Construction, Manufacturing, and Transportation and Warehousing is a straightforward benefit for workers compensation premium. These sectors combine to account for nearly half of premium in NCCI states.

While jobs in the Information, Finance and Insurance, Real Estate, and Educational Services sectors are typically “safer” jobs with lower loss costs, they still represent a significant share of premium due to the sheer volume of payroll in these sectors. Employment declines in lower-severity sectors combined with employment growth in high-severity sectors over an extended period of time could shift a larger share of premium away from these “safer” sectors.

AI’s Impact on Productivity

Productivity is often one of the earliest indicators of technological change. Measured as output per hour worked, labor productivity captures how efficiently the economy converts labor into economic value. Real output per hour is calculated by taking real gross domestic product (GDP, the quantity of output produced, with prices held constant) and dividing it by total labor hours worked. This calculation can be volatile in any given quarter due to changes in the growth of real GDP, particularly around recessions. Therefore, longer-term trends (patterns over several years rather than quarters) are more important to keep in mind.

Since mid-2023, productivity growth has been elevated. This trend of higher productivity growth comes after an extended period of low productivity growth experienced from the mid-2000s through the late 2010s. The more recent trend in productivity looks similar to the elevated trend experienced in the late 1990’s and early 2000s, around the time of the proliferation of personal computing and the Internet. While encouraging to see, questions remain around how long this new trend will persist. AI adoption has likely been a contributing factor and is likely to grow as a more important factor going forward, which may help sustain this trend of higher productivity.

A More Productive Workforce Is a Higher Paid Workforce

Higher productivity has several meaningful implications for the economy and workers compensation.

For the economy, the most straightforward benefit is a higher GDP for any given level of employment. With the U.S. economy facing constrained population growth and a constrained labor force, productivity will become an even more important driver of economic growth. Last year was a good example. In a year that saw barely any employment growth, the economy still experienced a solid 2% increase in GDP.

For workers compensation, productivity growth has an even more important impact: wages. Higher output per hour from workers typically leads to higher compensation per hour as firms compensate more productive workers. While wage growth has softened somewhat since its elevated levels coming out of the pandemic, the trend has remained solidly above the trend in the years prior to the pandemic.

Over the past two decades, outside of the volatile periods during and immediately following recessions, wage growth has tracked closely with productivity growth. At a time when other labor market indicators (slower hiring, lower job openings) would suggest softening wage growth, wages have continued to grow persistently above trend over the past several years, likely due to this trend of higher productivity.

This leads to an interesting chicken-or-egg question for the labor market: Has hiring slowed down because workers are more productive or have workers become more productive because hiring has slowed down and they must do the same work (or more) with fewer labor resources?

This is a topic that we intend to explore in more detail in a future brief as more information comes in. The answer has important implications for claim frequency trends. If the former proves to be the case, then frequency may decline faster as more productive workers, enhanced with technology, may be safer in completing tasks than in the past. Conversely, if the latter is true, then increased stress on workers could lead to fatigue, burnout, and more injuries.

Conclusion

It is still too early to assess the impact that AI is having on the labor market, and this story will continue to develop as technology evolves and adoption expands.

However, early signals are beginning to emerge. While adoption has been limited and uneven across industries, early 2026 data suggests a negative correlation between AI adoption rates and employment growth.

As we consider how AI will impact employment in the future, dimensionality remains a critical lens. Jobs with high dimensionality are more likely to see productivity gains and wage growth, while jobs with low dimensionality are more at risk of elimination and may require workers to evolve.

Higher productivity in recent years has supported both economic growth and wage growth. However, important questions remain about the cause of this higher trend—and the answer has meaningful implications for workplace safety and workers compensation frequency trends.