Artificial intelligence has moved past the promise stage and become a daily reality for businesses — and the ones feeling it the most are recent graduates who just stepped into the job market.
In Texas, two-thirds of surveyed companies are already using generative AI in their processes, a significant jump from the 40% recorded just two years ago, according to data from the May 2026 Texas Business Outlook Survey.
That number isn’t just a nice statistic for a report — it represents a quiet, profound shift in how companies hire, organize teams, and assign tasks.
And here’s a detail that makes this picture even more relevant: it’s not the digitally native startups driving this transformation.
It’s the traditional, well-established companies that are rethinking their hiring needs and, in many cases, simply opening fewer positions than before — especially for roles that AI can handle efficiently.
The question hanging in the air is: who’s paying the price for this transition? 🤔
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
The data points to a very specific — and vulnerable — group.
In this article, we’ll break down what an analysis based on millions of online job postings revealed about the impact of automation on the Texas labor market, which professions are on the front lines of this shift, and why recent graduates are being disproportionately affected.
How to measure a profession’s exposure to AI
Before understanding the effects, we need to talk about how economists actually measure all of this. The logic is interesting: researchers view any job as a bundle of tasks a worker performs regularly. Think about a medical records technician, for example. Their work involves compiling and maintaining patient records, documenting conditions and treatments, organizing and coding data using standardized systems, and reviewing records to ensure accuracy and regulatory compliance. Each of those activities is a piece that forms the puzzle of the profession.
The question that comes up is straightforward: how many of those pieces can generative artificial intelligence already put together on its own? To answer that, the analysis used a task-based metric developed by Anthropic, the company behind Claude, one of the leading AI assistants on the market. This metric cross-references the tasks listed in the O*NET database — which catalogs roughly a thousand common occupations in the American economy and details what each one does — with the tasks people actually perform using Claude on a day-to-day basis.
The brilliant part here is that this measure doesn’t evaluate what AI could theoretically do, but rather what it actually does cheaply and efficiently enough to be adopted by businesses and consumers. In the case of the medical records technician, for example, nearly 5% of the profession’s tasks are already automatable based on real-world usage of the tool. And which professions lead this exposure ranking? Generally, fields tied to software development, web design, and other computer-intensive occupations. But managers, administrative workers, editors, and other office roles also show up among the most exposed. 💻
What the data actually shows about employment and automation
To connect this exposure to labor market reality, the researchers used data from Lightcast, formerly Burning Glass, which collects and standardizes job postings from more than 220,000 job sites and company pages. This allows near real-time tracking of how labor demand across different professions evolves over time. An important caveat is worth noting: these online postings better represent certain types of work, and fields like agriculture, construction, and personal services are underrepresented in this data.
With that disclaimer out of the way, the results are eye-opening. After the launch of ChatGPT in late 2022 — a true watershed moment that triggered a wave of adoption and investment — job postings for professions more exposed to automation began declining consistently. Comparing more and less exposed occupations within the same industry, the drop was about 5% by the end of 2023 and reached roughly 8% in the first quarter of 2025. These percentages represent the difference associated with a ten-point variation in the share of automatable tasks within each role.
What makes this even more compelling is that the pattern holds when the analysis is expanded to the entire United States, with very similar results. In other words, this isn’t an isolated phenomenon or a one-off slowdown in a specific sector like tech. It’s a structural shift cutting across industries and regions. 📉
Traditional companies are hiring less
This is where one of the most surprising findings from the research comes in. You’d naturally assume that this drop in hiring was caused by the emergence of new companies born with AI baked in, needing fewer people. Or simply by a reduction in the number of active businesses. But the data tells a different story.
By analyzing a stable group of companies present throughout the entire study period, the researchers found that already established companies more exposed to AI reduced their hiring at rates very similar to the overall effect. The decline was roughly 5% to 6% through mid-2024 and reached 8% to 9% by early 2026. So it’s the traditional companies pumping the brakes — not just digital startups.
Beyond that, these companies changed the composition of the positions they offer. Companies whose roles were set to become 10% more automatable began posting jobs with two percentage points fewer automatable tasks after ChatGPT — a reduction of nearly 50% relative to the average. This shows that AI is acting as a substitute for certain tasks, and companies are reorganizing their teams to reflect this new reality. 🤯
Why recent graduates are hit the hardest
Based on these calculations, the researchers estimate that exposure to generative AI reduced total online job postings in Texas by about 1.8% in 2024 and 2.6% in 2025. At first glance, it looks like a modest effect in the aggregate. But when you look at specific groups of workers, the impact takes on a whole different dimension.
And there’s a harsh logic at play when you really think about it. Companies adopting artificial intelligence aren’t necessarily laying off experienced professionals in droves — at least not yet. What’s happening is a quiet reorganization: instead of bringing in new people, they simply hold off on hiring. Even without layoffs, this hiring freeze hits hard for anyone entering the job market for the first time or trying to make a career change.
The Lightcast numbers back this up. Fewer than half of typical job postings require more than two years of experience, and very few ask for more than five. Since most online listings are for roles that demand little prior experience, a decline in postings disproportionately affects newcomers — and nobody is more at the starting line than recent graduates.
It’s no surprise that the unemployment rate among recent graduates has climbed to unusual levels during exactly this period of accelerated AI adoption. There’s a phenomenon that has earned a name in research circles: entry-level gatekeeping. The gateway to the professional world is getting narrower, and the bar to clear it keeps rising. Entry-level positions historically didn’t exist just to get tasks done — they served to train professionals. That’s where someone learned how work actually functions. If those rungs disappear, today’s graduates are left without the ladder previous generations used to climb. 💡
What this landscape means for the years ahead
The question everyone is asking is whether this reduction in labor demand for recent graduates is temporary or structural. Administrative data from the state of Texas already suggests that automation has affected outcomes like employment and wages for recent graduates, and has even led current students to change their educational decisions to adapt to this new reality. In other words, the effect doesn’t just stay on paper — it’s changing people’s life choices.
Some argue that, as with other technological revolutions, new roles will emerge to replace the ones that vanish. Industrial automation created office work. Digitization spawned an entire digital services industry. Historical logic suggests the same could happen with AI — but with one important difference: the speed of this transformation is much faster, and the gap between old roles disappearing and new ones appearing could be long enough to affect an entire generation.
On top of that, the roles AI creates — like automation specialists, senior-level data analysts, and systems architects — require deep technical profiles and, often, prior experience. They don’t easily reach people who are just walking out of college. For young professionals, the message from the data is clear: the path into the job market is going to demand more than a diploma. It’s going to require hands-on familiarity with automation tools, the ability to work alongside AI instead of competing against it, and a skill set that sets them apart from what any language model can deliver on its own. It’s not a simple task — but it’s increasingly what will separate those who land an opportunity from those left waiting for the market to normalize.
