AI's Impact on Entry-Level Jobs: A Canary in the Coal Mine?

One year ago we discussed the idea about how AI is displacing jobs or will do in the future. The main result was: it depends on the type of job, specifically the tasks required for a job. Have a look at that article here. Now, a new paper has been published which tries to validate empirically this thesis. In the paper "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence", published by Standford Digital Economy Lab, the authors delve into high-frequency payroll data to uncover some potentially concerning trends in the labor market following the widespread adoption of generative AI tools. Instead of looking at overall numbers, which can hide important shifts, they drill down to look at specific age groups and specific industries.

The "Canaries" and What They're Telling Us:

The title of the paper refers to the historical practice of bringing canaries into coal mines to detect dangerous gases. If the canaries started showing distress, it was a warning sign for the miners. In this case, the authors present six "facts" that may be early warning signs about AI's impact, particularly on young workers in certain fields. Here's a breakdown of those findings:

Key Findings from the Stanford Digital Economy Lab Paper:

1. Entry-Level Employment Decline: The study found a significant drop in employment for workers aged 22-25 in occupations most exposed to AI, such as software development and customer service. This decline is relative to both more experienced workers in the same fields and to younger workers in less AI-exposed occupations.

   • Why this matters: This suggests AI isn't just affecting all jobs equally. It's potentially impacting the entry point for certain career paths.

     
     

2. Stagnant Growth for Young Workers: While overall employment is still growing, the research indicates that employment growth for young workers, especially those just starting out, has stagnated since late 2022. The paper suggests that this stagnation is due to the declines in employment AI-exposed jobs.

   • Why this matters: The economic health is masking trends for younger workers.

     
     

3. Automation vs. Augmentation Matters: AI's impact isn't uniform. The declines in entry-level employment are concentrated in areas where AI is primarily being used to automate tasks rather than augment human work. In occupations where AI usage is more complementary to human skills, employment growth is still observed.

   • Why this matters: This is a crucial distinction. The type of AI deployment heavily influences its impact on jobs.

     
     

4. Firm-Level Shocks Controlled: The authors took pains to rule out other potential explanations for the trends they observed. They controlled for firm-specific or industry-specific shocks (like interest rate changes) that might disproportionately affect younger workers in AI-exposed fields. Even after accounting for these factors, the relative employment decline for young, AI-exposed workers remains.

   • Why this matters: Adds confidence that the trends are linked to AI, not just other economic factors.

     
     

5. Employment More Affected Than Compensation (So Far): The labor market adjustments the study detected are primarily visible in employment figures, not in immediate compensation changes. The authors suggest this could be due to wage stickiness – wages are often slower to adjust than employment levels.

   • Why this matters: This suggests potential pressure on employment numbers as the labor market adjusts to AI.

     
     

6. Robust Across Scenarios: The authors tested their findings under a variety of alternative conditions:

   • Excluding technology-related occupations

   • Excluding occupations amenable to remote work (to address outsourcing concerns)

   • Looking at a longer time horizon (to understand pre-AI trends)

   • Analyzing occupations with both high and low college graduate shares

   The core findings remained largely consistent, strengthening the argument that AI is playing a role.

   • Why this matters: Higher confidence that this are really trends.

Finally, have a look at this figure taken from the article:

Employment trends for different age groups, showing a decline for young workers (22-25) in occupations with high AI exposure. Data from Stanford Digital Economy Lab study.

We can clearly see how, for younger workers, the employment trends for more AI-exposed occupations diverge downward compared to less-exposed ones. It also shows that the effect diminishes for older age groups, driving home the point that the impact is not uniform across all demographics.

Implications and What's Next:

The paper's findings suggest that the AI revolution is beginning to have a significant and disproportionate impact on entry-level workers in the American labor market. The findings are especially important for entry-level workers because the changes are noticed earlier. The article mentioned that AI is replacing codified knowledge, the "book-learning", the one obtained at formal education.These findings are just a snapshot in time, and the authors stress the need for ongoing monitoring to see how these trends evolve. It's crucial for companies to think carefully about how they deploy AI, focusing on augmentation and reskilling initiatives to help workers adapt to the changing landscape.