Stanford Digital Economy Lab’s AI Economic Indicators point to a sharp drop in employment for US workers aged 22 to 25 in jobs with high exposure to artificial intelligence. The most visible impact shows up in software development and customer service, where AI-driven automation and augmentation have started to reshape job requirements. Analysis of ADP payroll records found that, as of June 2026, employment among 22–25-year-olds in the most AI-exposed roles was about 19% below the level expected if their employment had matched the growth of peers in less-exposed jobs. The affected positions included software development and customer service, while comparison roles with less AI exposure included nurses’ aides and maintenance and repair workers. The employment gap centered on young workers. Older employees in similar AI-exposed jobs did not see the same decline.
AI exposure and employment trends
The Canaries Dashboard, built by Stanford Digital Economy Lab with ADP Research, tracks employment patterns using payroll data from a sample of US firms. The analysis covered millions of US workers and compared employment paths across occupational groups with different AI exposure levels. From November 2022 to June 2026, employment among 22–25-year-olds in the most exposed jobs reportedly fell about 11%. In contrast, employment in the least-exposed group rose about 10%. This gap reflects both contraction in highly exposed roles and growth in lower-exposure jobs. The analysis separates AI automation, where tasks shift to systems, from augmentation, where workers use AI alongside their own skills. Automation-heavy jobs showed the clearest negative employment link for early-career workers. This pattern did not appear as strongly in older age groups, where employment in highly AI-exposed roles stayed steady or even grew.
Software developers and customer service representatives in the youngest group are among those most affected. Older workers in these roles have seen employment expand. Meanwhile, lower-exposure jobs such as nurses’ aides and maintenance and repair workers have posted stronger employment gains among young workers. These findings come from Stanford Digital Economy Lab’s Canaries in the Coal Mine? research program, which tracks whether labor-market effects show up first among younger and less-established workers. The August 2026 update extended the payroll analysis through June 2026, letting researchers compare the post-November-2022 period with earlier employment patterns.
Understanding AI’s role in workforce changes
The dashboard also looks at how AI is used within different jobs. It distinguishes between automation, where tasks are fully handed off to AI, and augmentation, where people work with AI systems. Jobs with more automation tend to show employment declines or limited growth, especially for early-career workers. The link between augmentation and employment trends is less clear for this group. The Stanford results do not claim that AI eliminated a set number of jobs. Instead, they show a split in employment outcomes between young workers in exposed and less-exposed jobs. The available reporting points to weaker hiring, not a spike in layoffs, as the main reason for the employment gap. Education plays a major role: controlling for occupational education requirements roughly halved the estimated gap and, in one sample, erased statistical significance.
Calls for new learning models
Isabelle Hau, Executive Director of the Stanford Accelerator for Learning, highlighted these findings on LinkedIn and stressed the need for new approaches to professional development. She wrote, “Entry-level jobs have traditionally been essential environments for learning by doing, developing judgment, building relationships, and growing into professional roles.” Hau called for new models of apprenticeships, mentorship, and experiential learning to help young people build expertise as AI transforms the workplace. She also argued that human capabilities such as relational intelligence, creativity, critical thinking, agency, and adaptability should become explicit learning outcomes, rather than by-products of education.
Limitations and context
The Canaries Dashboard measures correlations between occupational AI exposure and employment trends but does not prove that AI directly caused the observed declines. As Stanford researchers summarized, “The data show a pattern, not a cause.” The underlying economic mechanism remains debated because young workers are overrepresented in entry-level hiring pipelines, where firms may cut recruitment, training, or supervision even without large layoffs. Australian Assistant Minister for Competition, Charities, Treasury and Employment Andrew Leigh has cited research suggesting that remote work may also make it harder to supervise and train new hires, offering a non-AI explanation for part of the youth hiring slowdown. The dashboard uses the Anthropic Economic Index to assess how AI is used within jobs and is updated monthly, with the latest update on October 21, 2026. Alongside the Canaries Dashboard, the AI Economic Indicators include tools for tracking AI adoption, economic transformation, and consumer surplus from generative AI tools. The Transformation Tracker, for example, monitors output growth, labor productivity, and capital shares. As of October 21, 2026, it reports no decisive evidence of a shift toward explosive economic growth.
The latest Stanford update puts the relative employment gap for 22–25-year-olds at about 19% through June 2026, compared with an earlier estimate of about 15% in the initial version of the study published in August 2025. The evidence does not show broad economy-wide displacement. The clearest effect remains weaker employment and hiring outcomes for young workers in highly exposed jobs, while older workers have not seen a similar decline. These findings highlight the need to adapt educational and professional development models so young workers can succeed as AI continues to reshape the labor market. According to the Stanford Digital Economy Lab, ongoing updates to the dashboards are publicly accessible and provide valuable insights for educators, policymakers, and employers navigating the evolving landscape of AI in the workforce. For further reading, see the Fortune report on AI and employment.
