Archive · 25 Aug 2026
Stanford Study Finds AI-Exposed Entry-Level Jobs Falling 19 Percent
A Stanford study looked at how artificial intelligence may be changing jobs.
It found that people aged 22 to 25 had fewer jobs in occupations considered highly exposed to AI.
Their employment was 19 percent lower than for young workers in less-exposed occupations.
The main difference appeared to come from fewer entry-level hires, not lower pay.
Across all age groups, the researchers found little difference in total employment between the two types of jobs.
Jobs based mostly on written rules and information may be easier for AI to perform.
Jobs requiring practice, mentoring, and real-world experience may work better with AI helping people.
The researchers said college education may reduce, but not eliminate, these differences.
Employment for workers aged 22 to 25 in the most AI-exposed occupations was 19 percent lower than in less-exposed fields.
The Stanford researchers linked the gap mainly to reduced hiring, rather than lower wages, for entry-level workers.
Among young workers, employment fell about 11 percent in the 40 percent of jobs most affected by AI, while it grew 10 percent in the least-exposed 60 percent.
Overall employment showed little relative difference between the most and least AI-exposed occupations.
The study found that jobs based on codified knowledge may be more vulnerable, while roles requiring practice and mentorship may be more likely to be complemented by AI.
- Who
- Economists at Stanford University, led by Erik Brynjolfsson, studied employment among workers, particularly those aged 22 to 25.
- What
- The study found that entry-level employment has declined more in occupations considered highly exposed to AI.
- When
- The article describes an August 2026 study and compares its findings with employment trends identified the previous year.
- Where
- The analysis examined employment across occupations using anonymized payroll data; the article does not identify a specific geographic location for the data.
- Why
- Researchers attributed the difference mainly to lower hiring in AI-exposed occupations and examined which types of work are more vulnerable or complementary to AI.
AI Is Closing the Entry-Level On-Ramp
AI May Reshape Jobs Rather Than Eliminate Them
Effect on young workers
AI Is Closing the Entry-Level On-Ramp
The study found that employment for 22- to 25-year-olds is substantially lower in the most AI-exposed occupations, primarily because of reduced hiring.
AI May Reshape Jobs Rather Than Eliminate Them
The study found little difference in overall relative employment between the most and least AI-exposed occupations when all age groups were considered.
Which jobs are vulnerable
AI Is Closing the Entry-Level On-Ramp
Jobs involving heavily codified knowledge that can be taught through textbooks or written procedures may be more vulnerable to replacement.
AI May Reshape Jobs Rather Than Eliminate Them
Jobs requiring practice, mentorship, and repeated exposure to real situations may be complemented by AI rather than replaced.
Role of education
AI Is Closing the Entry-Level On-Ramp
Young college graduates may face a shrinking entry-level pathway in some AI-exposed fields.
AI May Reshape Jobs Rather Than Eliminate Them
The study found that occupations with a higher share of college graduates had more muted employment differences between AI-exposed and less-exposed jobs, suggesting education may provide some protection.
- Study title
- “Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence.”
- Researchers
- Economists at Stanford University, led by Erik Brynjolfsson.
- Young-worker age range
- 22 to 25 years.
- Employment gap
- Employment in the most AI-exposed occupations was 19 percent below that in less-exposed occupations for this age group.
- Most-exposed occupations
- Employment among young workers fell about 11 percent in the top 40 percent of AI-impacted jobs.
- Least-exposed occupations
- Employment among young workers grew 10 percent in the 60 percent of jobs least affected by AI.
- Data source
- A large anonymized, high-frequency payroll-data sample aggregated by ADP.
- Education finding
- Occupations with more college graduates showed more muted differences between AI-exposed and less-exposed jobs.