On October 1, 2025, researchers Martha Gimbel, Molly Kinder, Joshua Kendall and Maddie Lee released one of the most anticipated labor reports of the year: Evaluating the Impact of AI on the Labor Market.
The findings may surprise many who have been bracing for a wave of job losses triggered by generative AI since the release of ChatGPT nearly three years ago.
The conclusion? Despite all the headlines about automation taking over, the broader U.S. labor market looks remarkably stable.
There is no discernible evidence, at least yet, that AI is driving widespread job loss or unemployment. But what does this really mean and should we worry about the future?
Has AI Changed the Labor Market Since 2022?
The big question driving this study was straightforward: Has generative AI disrupted the labor market since its public debut in November 2022?
The researchers looked at 33 months of employment data, a period marked by global adoption of AI tools across industries and compared it with previous waves of technological disruption.
It also includes the rise of computers in the 1980s and the internet boom in the 1990s.
The data shows that the occupational mix is indeed shifting but not in ways that are unprecedented.
For example, by 2002, six years into the internet era, the occupational mix had only shifted by about 7 percentage points from 1996.
Today, after nearly three years of AI adoption, changes in the job mix look very similar, about one percentage point higher, but still within historical norms.
So far, AI’s effects do not look out of the ordinary compared to past technological revolutions. As the report notes, “technological disruption in workplaces tends to occur over decades, not months.”
Is This Time Different From the Past?
AI certainly feels different. It is faster, more versatile and more accessible than past innovations. But when it comes to labor market disruption, is it actually moving faster?
The researchers’ analysis of the dissimilarity index is a measure of how different today’s job mix is from a past baseline suggests that AI’s effect is not dramatically different from computers or the internet.
Yes, jobs are shifting but the scale is modest.
And perhaps most telling: much of the occupational reshuffling we’re seeing actually began before AI tools were released.
By 2021, before ChatGPT, industries like Information and Financial Activities were already showing changes in job composition. AI may be a factor, but it is not the sole driver.
This is a reminder that labor markets are always evolving. Sometimes we attribute too much to one technology without considering broader economic forces.
Which Industries Are Most Affected by AI Exposure?
The report highlights three industries showing above-average change since 2022:
- Information (including data processing, media, and publishing)
- Financial Activities
- Professional and Business Services
These industries also happen to be the most exposed to AI augmentation and automation. For instance, writers, analysts and customer service roles face clear overlaps with AI’s capabilities.
But again, the twist: these industries were already volatile before AI entered the picture.
In fact, the Information sector has been undergoing dramatic shifts for nearly two decades, from the digitization of newspapers to the rise of streaming. AI is just the latest in a line of forces reshaping it.
What About Young Workers and Recent Graduates?
Much of the anxiety about AI centers on early career workers. After all, entry-level roles are often the first to be automated.
So is AI making it harder for recent graduates to break into the labor market?
The data here is nuanced. Comparing recent grads (ages 20–24) to slightly older ones (25–34), the researchers found only slight differences in occupational mix since 2022.
Yes, dissimilarity has risen a little faster in recent months, which could suggest younger workers are being nudged into different roles.
But the trend could also reflect a generally slower labor market recovery.
In short, AI might be influencing young workers’ job paths but not yet in a way that shows up clearly at the economy-wide level.
What Do “Exposure” and “Usage” Tell Us?
One of the most interesting parts of this report comes from comparing exposure data (from OpenAI) with usage data (from Anthropic’s Claude).
- Exposure measures whether AI could reduce the time it takes to do a task by 50% or more.
- Usage looks at how AI is actually being applied in real workplaces.
OpenAI’s data shows that roughly 18% of workers are in highly exposed occupations—jobs where AI could handle a significant chunk of tasks.
But the share of workers in these jobs has not meaningfully shifted since 2022.
Anthropic’s usage data tells a similar story. Occupations with high actual AI use such as coding, writing, data analysis are stable.
Interestingly, AI usage is heavily concentrated in computer science and media roles, while clerical and production roles, despite being theoretically “exposed,” show much lower usage.
The mismatch between exposure and usage highlights an important truth: just because AI can do a job does not mean it is being used that way and at least not yet.
Why Haven’t We Seen Major Job Losses?
So why, despite all the hype, has not AI triggered massive layoffs?
There are three key reasons:
- Technology Diffusion Takes Time
Computers took nearly a decade to reshape office workflows. The internet took even longer to transform commerce. AI will likely follow a similar timeline. - AI Often Augments, Not Automates
Many workers are using AI to make tasks faster or easier, not to eliminate their role entirely. For example, coders use AI for debugging, but they are not being replaced wholesale. - Economic Forces Beyond AI
Shifts in industries like finance or information predate AI. Broader market conditions, regulation, and consumer behavior still matter as much as technology.
Do We Need Better Data to Understand AI’s Impact?
The authors are clear: today’s data is not sufficient to fully capture AI’s labor market impact. OpenAI’s exposure metrics are theoretical, while Anthropic’s usage data is limited to one AI tool and a small subset of occupations.
As the report urges, more transparent, comprehensive usage data is needed from AI companies. Without it, researchers are left with proxies that may understate or misstate the true effects.
This is especially important as AI adoption spreads into new areas like manufacturing, logistics and healthcare. What looks stable today could mask deeper undercurrents of change.
So, Is AI Disrupting the Job Market Yet?
The short answer: not in any measurable, economy-wide way. The U.S. labor market has seen no spike in unemployment, no collapse in job categories and no structural shift on par with the automation fears dominating headlines.
But the long answer is more cautious: we may simply be too early. AI’s potential is vast, and history tells us that technological disruption plays out over decades. The anxiety isn’t misplaced but the timeline is longer than many assume.
As the authors put it, “The picture of AI’s impact on the labor market that emerges from our data is one that largely reflects stability, not major disruption at an economy-wide level.”
Dileep Thekkethil
AuthorDileep Thekkethil is the Director of Marketing at Stan Ventures, where he applies over 15 years of SEO and digital marketing expertise to drive growth and authority. A former journalist with six years of experience, he combines strategic storytelling with technical know-how to help brands navigate the shift toward AI-driven search and generative engines. Dileep is a strong advocate for Google’s EEAT standards, regularly sharing real-world use cases and scenarios to demystify complex marketing trends. He is an avid gardener of tropical fruits, a motor enthusiast, and a dedicated caretaker of his pair of cockatiels.