AI Is Shrinking Entry-Level Jobs in Tech and Accounting

What does the new Stanford and ADP data show about young workers?

Employment among workers aged 22 to 25 in AI-exposed occupations fell 4.2% in April 2026 compared to a year earlier, according to new data from the Stanford Digital Economy Lab and ADP. Workers of similar ages in less-exposed fields saw a smaller 1.7% decline over the same period.

The broader picture looks calmer. Employment in AI-exposed jobs across all ages dropped just 0.2% in April year-over-year, while employment in less-exposed fields grew 0.1%. The damage is concentrated at the start of careers.

Which jobs are hit hardest?

Software development and customer service are the clearest examples. After peaking during the strong labor market of 2022, employment in software development for early-career workers began falling after the launch of ChatGPT. Workers older than 30 in the same field kept seeing employment growth over that period.

Stock clerks, considered less exposed to AI, did not follow the same pattern. Home-health aides, a fast-growing field, saw the opposite trend — young workers there saw the biggest employment surge since ChatGPT launched.

How does the Stanford "Canaries Dashboard" track this?

The Canaries Dashboard is a monthly employment tracker built from anonymized ADP payroll data, run as a collaboration between ADP Research and the Stanford Digital Economy Lab. It covers roughly 4.6 million workers across 25,000 firms.

The dashboard takes its name from a November 2025 Stanford paper titled "Canaries in the Coal Mine," which identified early-career workers in AI-exposed fields as the first group to show employment setbacks after generative AI became widespread.

You might also like

US Deficit Projected at $2 Trillion in FY2026

TSMC CEO Rules Out Memory-Style Price Spikes

Google Signs Secret Pentagon AI Deal, Director Quits

71% of Americans Oppose Local AI Data Centers

ADP chief economist Nela Richardson told Yahoo Finance the dataset shows why broad labor statistics miss the story. "AI hits at the task level," Richardson said. "You really have to have the microscope ready on data."

Early-career workers account for about 7% of ADP's sample, Richardson noted, which is why aggregate numbers mask the trend.

What is happening to entry-level accounting jobs?

A BambooHR survey published May 12, 2026 found a 3:1 senior-to-entry-level hiring ratio in accounting and finance at small to midsize U.S. businesses. One-third of new accounting and finance hires quit within their first year.

The survey collected 1,248 responses from full-time salaried employees, business owners, and C-suite leaders between March 24 and April 9, 2026. BambooHR also drew on six years of workforce data covering over 480,000 employee data points from more than 2,000 companies worldwide.

Why are senior accountants staying while entry-level roles shrink?

BambooHR CFO Justin Judd said senior tools are absorbing what junior staff used to do. "Senior-level accountants and senior-level analysts in finance have new tools where they can do some of the work that used to be done by entry-level team members on their own," Judd told CFO Brew.

Historically, entry-level accounting roles had a clear scope: data entry, spreadsheet building, database work. That clarity is gone. "What that leaves is a question of, 'What do I do in an entry level role?'" Judd said.

The result is that companies are hiring more experienced workers and keeping senior staff longer. "With the new tools, what do I need? I actually need someone with judgment still, the ability to discern," Judd said.

What skills do employers now expect from entry-level hires?

In CFO Brew's State of the Industry report, 77% of respondents named strategic thinking and problem solving as the most important skills for the next generation in finance and accounting.

Judd described the new expectation as architectural rather than executional. "You are an architect of a system, you're an architect of a process where, in the past, you might have just been doing work; now you're orchestrating or architecting automation," he said.

Can early-career workers in AI-exposed fields recover?

Richardson said the data should push young workers to act, not retreat. "The issue is not changing jobs or changing occupations," she said. "The issue is how to advance your skillset within this occupation to jump from that automation rung to the augmentation rung. And that is completely possible."

The Stanford Digital Economy Lab noted that AI-based augmentation does not produce the same employment declines seen with automation — and that the negative pattern "disappears as you move from automation to augmentation, and as you get increasingly more complex jobs."

Judd pointed to structured onboarding as one concrete response. BambooHR is "currently spending a lot of time making sure folks are trained, they have access to the tools, they're being thoughtful about what it means," he said.