In March 2026, the Federal Reserve Bank of Dallas released a study that exposed a stark new reality in the U.S. Labor market: artificial intelligence is not just reshaping jobs—it is creating a two-tiered economy where experience is the new currency, and entry-level workers are being left behind. The findings, based on wage and employment data from late 2022 to early 2026, reveal a workforce divide so pronounced that it challenges decades of conventional career progression. While senior professionals in AI-exposed sectors are seeing wage growth outpace national averages by nearly double, young workers—particularly those in technology, law, and finance—are facing shrinking opportunities, with some roles effectively disappearing before they can even be filled.
The study’s most striking conclusion is that AI’s impact is not uniform. Instead, it is amplifying an existing disparity between two types of knowledge: codified and tacit. Codified knowledge—the kind taught in classrooms, documented in manuals, and now instantly replicable by AI—is being devalued. Tacit knowledge, the intuitive expertise gained through years of on-the-job experience, is becoming more valuable than ever. “AI excels at processing information and applying rules, but it cannot replicate the judgment of a seasoned lawyer sensing a deal’s collapse or an engineer anticipating system failure before the data confirms it,” said Heather McGowan, a labor market analyst at the University of Southern California, whose research aligns with the Dallas Fed’s findings. “This is the core of the divide—and it’s showing up in paychecks.”
The wage data tells a compelling story. Since the launch of ChatGPT in November 2022, nominal average weekly wages across the U.S. Have risen by 7.5%, according to the Bureau of Labor Statistics. But in the computer systems design sector, one of the most AI-exposed industries, wages have surged by 16.7% over the same period. Among the top 10% of AI-exposed industries—including legal services, financial analysis, and marketing—the average wage growth has been 8.5%, far outpacing the national rate. The disparity is even more pronounced when broken down by age. Workers aged 22 to 25 in AI-exposed roles have seen employment drop by 16% since late 2022, while their older colleagues remain largely unaffected. In software development, junior roles for this age group have declined by 20% from their peak in late 2022, according to Cornell University’s AI Hiring Impact Study.
The consequences extend beyond individual paychecks. For generations, the entry-level job was the first step on a ladder that, with time and effort, could lead to the C-suite. Stories like Doug McMillon’s rise from Walmart truck unloader to CEO or Mary Barra’s ascent from GM assembly line worker to CEO were once emblematic of the American workforce. But those pathways are now collapsing. SignalFire, a venture capital firm tracking hiring trends at major tech companies, found a 50% decline in new role starts for people with less than one year of post-graduate experience between 2019 and 2024. The drop was consistent across sales, marketing, engineering, operations, finance, and legal, suggesting this is not a sector-specific issue but a structural shift.
“The loss of clear entry points doesn’t just shrink opportunities for new graduates—it reshapes how organizations grow talent from within,” said Heather Doshay, a partner at SignalFire. “Companies are no longer willing to invest in potential. They want proven experience.”
The data on job postings reinforces this trend. Entry-level roles requiring zero to two years of experience have dropped by 29 percentage points since 2022, according to DesignRush’s AI Job Displacement Statistics 2026. In the software and IT sectors, more than 60% of so-called entry-level roles now demand three or more years of experience, creating a paradox where fresh graduates are effectively barred from the first rung of the ladder. Meanwhile, healthcare—one of the few sectors still actively hiring entry-level workers—has seen postings rise by 13 percentage points, as clinical judgment and patient interaction remain resistant to automation.
Not all experienced workers are benefiting equally. The Dallas Fed study identified occupations where the experience premium—the gap between entry-level and senior wages—is highest and where AI exposure is most intense. These include:
- Lawyers and legal professionals: Experience premium exceeds 100%, with AI augmenting research and drafting while senior judgment remains irreplaceable.
- Insurance underwriters and credit analysts: Complex risk assessment built on years of pattern recognition commands growing premiums.
- Marketing specialists: Strategic brand thinking and client relationships carry enormous tacit value that AI tools amplify rather than replace.
- Senior software engineers: Those who understand system architecture and business context are seeing wages soar even as junior coding roles shrink.
- Healthcare professionals: Clinical experience and patient judgment remain deeply human, with entry-level postings rising.
At the opposite end of the spectrum, roles where both entry-level and experienced workers perform codifiable tasks—such as fast-food preparation, ticket processing, or dry cleaning—are experiencing negative wage growth, as AI can substitute for workers at every level of seniority. This bifurcation is not just economic; it is reshaping the social contract of work itself.
The warnings from within the AI industry have been blunt. In March 2025 testimony before Congress, Dario Amodei, CEO of Anthropic, predicted that AI could eliminate roughly 50% of all entry-level white-collar jobs within five years, potentially pushing U.S. Unemployment rates to levels not seen since the Great Depression. He described the potential outcome as a “white-collar bloodbath” and urged lawmakers to act urgently. Boris Cherny, the Anthropic engineer behind Claude Code, went further, suggesting that the title “software engineer”—once the most reliable entry-level position in technology—could be effectively extinct by the end of 2026. Cherny noted in a 2025 internal memo that he had not written a single line of code himself since November 2025, having handed all coding tasks over to AI.
Already in the first two months of 2026, roughly 32,000 job losses were recorded in technology firms alone, according to Challenger, Gray & Christmas. In 2025, nearly 55,000 job cuts were directly attributed to AI by company announcements, out of a total 1.17 million layoffs across the U.S. Economy—the highest figure since the pandemic year of 2020.
Not all economists share the most alarmist forecasts. Anders Humlum, a labor economist at Goldman Sachs Research, argues that two and a half years of widespread AI adoption have not yet produced the kind of mass unemployment some predicted. Historical patterns suggest that even the most transformative technologies—steam power, electricity, computing—took decades to generate large-scale employment effects. “Even if AI eventually matches the technical capabilities of entry-level white-collar workers,” Humlum said in a March 2026 interview with The Wall Street Journal, “the time required for workflow adjustments and human adaptation is routinely underestimated.” Goldman Sachs projects that AI’s impact on overall employment will be relatively mild and short-lived, with unemployment rising by around 0.5% during the transition—a reflection of short-term friction rather than permanent structural collapse.
Yet the immediate reality for young workers remains tricky. Career experts warn that waiting for the job market to return to normal is not a viable strategy. Instead, they recommend:
- Master AI tools immediately: Employers increasingly expect new hires to arrive proficient with tools like ChatGPT, Copilot, and Claude. Treat AI fluency as a basic workplace skill, akin to Excel proficiency for previous generations.
- Build experience aggressively: Even unpaid internships, freelance projects, or volunteer work counts. In a market that no longer hires on potential alone, real-world experience sets candidates apart.
- Target sectors still hiring at entry level: Healthcare, skilled trades, government, and physical services remain active recruiters and are largely resistant to automation.
- Go narrow, not broad: Specialist knowledge in niche areas—such as AI ethics, cybersecurity, or green energy—is harder for AI to replicate than general business skills.
- Think about the experience premium: Before choosing a career path, assess whether the experienced version of that role commands significantly more than the entry-level version. High experience premiums signal long-term security.
A new model is emerging at the entry level: “AI Apprenticeships”, where junior workers use AI tools to perform at a mid-level capacity far earlier in their careers than previously possible. Universities are partnering with AI companies like Anthropic and OpenAI to train students to harness these tools from day one. However, the transition is not without risks. The National Association of Colleges and Employers (NACE) reported in its Class of 2026 Graduate Employment Report that only 38% of new graduates feel adequately prepared to integrate AI into their work, a figure that drops to 25% in technical fields.
The Federal Reserve’s study concludes with a stark observation: returns on job experience are increasing in AI-exposed occupations. In other words, the premium on having actually done something—rather than just having studied it—has never been higher. The career ladder is not gone. But for millions of young workers, the first rung is now much harder to reach.
