The 22-Year-Old Developer's Vanishing Job — What 4 Independent Studies Say, and the Counterargument

Four Studies, One Signal

When one study finds a troubling trend, it might be noise. When four independent research groups, using different data sources and different methodologies, converge on the same finding — that is a signal worth taking seriously.

In early 2026, four separate analyses all pointed to the same conclusion: entry-level jobs in AI-exposed occupations are declining, and young workers are bearing the brunt.

But the story is not as simple as "AI is killing junior jobs." One of these four sources provides a compelling counterargument that complicates the narrative considerably. Here is what each study found, and why the counterargument matters as much as the trend.

Source 1: ADP Research — The 6% Drop

ADP Research, working with Stanford researchers, analyzed payroll data covering millions of U.S. workers. Their finding: employment among 22-to-25-year-olds in AI-highly-exposed occupations dropped by 6% between late 2022 and mid-2025. ADP/Stanford Research

In the same period, workers aged 30 and older in those same occupations saw employment grow by 6-13%. ADP/Stanford

That divergence is striking. The same jobs — software developers, data analysts, financial associates — are growing for experienced workers while shrinking for newcomers. The implication is that AI is not eliminating these occupations wholesale. It is raising the bar for entry while making experienced workers more productive and therefore more valuable.

For a 22-year-old software developer entering the job market, this is a concrete economic headwind that did not exist three years ago.

Source 2: Dallas Federal Reserve — The Dual Effect

The Dallas Fed published two relevant analyses in early 2026. The first, from February 24, examines wage and employment data across AI-exposed occupations. The second, from January 6, focuses specifically on young workers. Dallas Fed, February 2026 Dallas Fed, January 2026

Their finding: AI is simultaneously aiding and replacing workers — and the effect splits along experience lines. Dallas Fed Occupations requiring experience-based tacit knowledge (knowledge that comes from years of practice rather than textbook learning) are seeing wage increases. Computer systems design jobs, for instance, saw wages rise 16.7% since fall 2022 even as employment growth in the sector lagged. Dallas Fed

Meanwhile, the share of employment in AI-highly-exposed occupations dropped from 16.4% to 15.5% of total U.S. employment. Dallas Fed The decline concentrated among younger workers. Dallas Fed, January 2026

Only 10% of firms surveyed by the Dallas Fed reported that AI had actually reduced their employment. Dallas Fed Most of those reductions were in customer service and process automation roles. The majority of firms reported using AI to augment existing workers rather than replace them — but "augment existing workers" in practice often means "do more with fewer new hires."

Source 3: HBR — IBM's Counterintuitive Move

Harvard Business Review's March 2026 piece on AI and entry-level jobs provides the most nuanced look. The article documents how most companies are reducing junior hiring as AI handles tasks that traditionally trained new employees — research, data compilation, basic analysis, first-draft writing. HBR, March 2026

But IBM's HR leadership offers a striking counter-example. Instead of cutting entry-level positions, IBM tripled its entry-level hiring. HBR, "AI and the Entry-Level Job," March 2026 Their reasoning: the nature of entry-level work is changing, not disappearing. New hires now spend less time on routine tasks and more time learning to work with AI systems — a skill set that IBM considers more valuable long-term than traditional junior-analyst grunt work.

This is an important data point because it suggests the entry-level decline is not inevitable. It may be a choice that most companies are making — and some, like IBM, are choosing differently.

Source 4: EIG — The Counterargument

The Economic Innovation Group (EIG) published what may be the most important piece of this puzzle in January 2026. Using Lightcast job posting data, they confirmed that job postings in AI-highly-exposed occupations peaked in March-April 2022 and then declined sharply. EIG, January 2026

Here is the critical detail: that decline began before the release of ChatGPT in November 2022. EIG

EIG argues that the entry-level job decline in tech and related fields is primarily driven by the interest rate cycle, not generative AI. When the Federal Reserve began raising rates in March 2022, tech hiring — which had been on a pandemic-era surge — contracted sharply. The timing overlap with generative AI's emergence is largely coincidental. EIG

The EIG paper also cites Danish labor market data showing that AI's impact on income and working hours has been "precise zeros" — statistically no measurable effect. EIG, citing Danish economic data

This does not mean AI has no effect on entry-level employment. But it does mean that separating AI's impact from the broader macroeconomic cycle — rising interest rates, post-pandemic normalization, a cooling tech sector — is extremely difficult. Some of the 6% decline that ADP measured may have happened regardless of AI.

What the Cross-Analysis Reveals

Taking all four sources together, the picture is more complex than either the "AI is destroying jobs" or "AI is fine" narratives suggest.

What the evidence supports: young workers in AI-exposed occupations face measurably worse employment outcomes than older workers in the same fields. The gap is real and large enough to affect career decisions.

What the evidence complicates: how much of this decline is caused by AI specifically, versus broader economic forces. EIG counterargument The interest rate cycle, post-pandemic hiring corrections, and a general tech slowdown all contribute. AI may be accelerating a trend that was already underway, rather than creating it.

What one company demonstrates: the decline is not predetermined. IBM's decision to triple entry-level hiring shows that companies can choose to invest in junior talent development rather than eliminate it. HBR

For a 22-year-old entering the job market today, the practical implication is clear: AI skills are no longer a differentiator for entry-level developer positions — they are a prerequisite. The employers who are still hiring juniors are looking for people who can work alongside AI from day one.

Explore how AI affects these roles: Software Developers, Data Entry Keyers, Customer Service Representatives.

Sources

• Dallas Federal Reserve. (2026, February 24). AI's Dual Impact on Workers: Aiding and Replacing. dallasfed.org
• Dallas Federal Reserve. (2026, January 6). AI and Young Workers. dallasfed.org
• ADP Research Institute & Stanford HAI. (2025). Assessing the Real Impact of Automation on Jobs. hai.stanford.edu
• Economic Innovation Group. (2026, January). AI and Jobs: The Final Word? eig.org
• Harvard Business Review. (2026, March). AI and the Entry-Level Job. hbr.org

Update History

• 2026-03-20: Added source links and ## Sources section
• 2026-03-17: Initial publication based on cross-analysis of Dallas Fed (Feb 2026), ADP/Stanford Research, EIG (Jan 2026), and HBR (Mar 2026)

---

This article was researched and written with AI assistance using Claude (Anthropic). Analysis synthesizes findings from four independent sources: Dallas Federal Reserve (February and January 2026), ADP Research Institute/Stanford, Economic Innovation Group (January 2026), and Harvard Business Review (March 2026). This is AI-generated analysis of publicly available research and should not be taken as professional career or employment advice. We encourage readers to consult the original sources linked throughout this article.