Insights · March 10th, 2026
Erik Brynjolfsson, Bharat Chandar and Ruyu Chen just released the paper ‘Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence‘ that analyzes high-frequency administrative data from ADP to understand how generative AI is reshaping the U.S. labor market. The authors argue that early-career workers serve as “canaries,” providing early evidence of AI’s labor market impact.
The abstract is as follows – “Using high-frequency administrative data from ADP, we document six facts characterizing labor market shifts following the widespread adoption of generative AI. Early-career workers (ages 22-25) in AI-exposed occupations experienced 16% relative employment declines, controlling for firm-level shocks, while employment for experienced workers remained stable. Adjustments occur primarily via employment rather than compensation, with employment changes concentrated in occupations where AI automates rather than augments labor. Results are robust to excluding technology firms and occupations that can be undertaken remotely. These six facts provide early large-scale evidence consistent with generative AI disproportionately impacting entry-level workers in the American labor market.”
Six Key Facts
1. Significant Employment Declines for Young Workers: Early-career workers (ages 22–25) in occupations most exposed to AI—such as software developers and customer service representatives—experienced a 16% relative employment decline. For example, employment for software developers in this age group dropped nearly 20% from its late-2022 peak.
2. Stagnant Overall Growth for Entry-Level Roles: While the broader U.S. labor market has remained robust, employment growth for workers aged 22–25 has been stagnant since late 2022. This stagnation is driven specifically by the 6% decline in AI-exposed jobs for this demographic, whereas older workers in the same roles saw growth of 6–9%.
3. Automation and the rise of Augmentation: Employment outcomes depend on how AI is used. Entry-level employment has declined in roles where AI primarily automates tasks (substituting for labor), but remained stable or grew in occupations where AI augments work (complementing labor).
4. Robustness to Firm-Level Shocks: The findings are not simply a result of broad tech industry struggles or interest rate changes. Even when controlling for firm-time effects (which account for shocks affecting all workers at a specific company), the relative decline for young workers in AI-exposed roles persists.
5. Adjustment via Employment, Not Wages: The labor market is adjusting through hiring volume rather than pay cuts. Salary trends for young workers in AI-exposed roles have remained stable, suggesting wage stickiness where firms choose to hire fewer new people instead of reducing compensation.
6. Consistency Across Sub-Groups: The trends hold even after excluding technology firms, computer-specific occupations, or roles susceptible to remote work. They also appear in both high-education and low-education occupations, suggesting that post-pandemic shifts in education quality are not the primary driver.
Where young workers are impacted most – tacit vs. codified knowledge
The authors propose that AI is highly effective at replacing codified knowledge—the “book-learning” and rules-based tasks typically mastered early in a career. Conversely, AI is currently less capable of replacing tacit knowledge, which includes the judgment, client management, and process-intensive skills that experienced workers accumulate over years on the job. Consequently, firms may be shrinking junior hiring (“task substitution at the apprentice margin”) while maintaining or increasing their reliance on experts.
2nd and 3rd Order Effects beyond the bottom line
While the immediate “1st order” effect is a reduction in payroll costs for junior roles, the long-term implications for your organization are far more complex.
2nd Order Effects: The Organizational Strain
- Mentorship Void: If you stop hiring juniors, your senior leaders lose the opportunity to develop management skills. You aren’t just losing “doers”; you are failing to train your next generation of “leaders.”
- Talent Premium Spike: As the supply of “proven” mid-level talent dries up (because fewer people are entering the pipeline today), the cost of hiring experienced experts will skyrocket. You are trading low-cost junior labor today for an unaffordable talent war tomorrow.
- Knowledge Concentration and Loss: Wisdom becomes siloed in a shrinking pool of aging experts. If your “tacit knowledge” holders retire without a junior class to shadow them, that institutional memory vanishes forever.
3rd Order Effects: The Systemic Shift
- Expert Paradox: If AI does all the entry-level work, how does a human ever gain the 10,000 hours of practice required to become an expert? We risk a future where we have “God-like” AI and “Novice” humans, with no bridge in between.
- Institutional De-skilling: As organizations rely on AI for “codified” tasks, the internal muscle memory for how those tasks are actually performed atrophies. If the AI fails or hallucinates, your organization may no longer have the human capacity to audit the work.
- The Higher Education Re-alignment: We are heading toward a massive “mismatch” crisis where universities continue to churn out graduates with “codified” skills that the market no longer values, leading to social friction and a potential “lost generation” of white-collar professionals.
From Automation to Augmentation
The Stanford research highlights a critical path forward: Augmentation over Automation. Employment remained stable or grew in roles where AI complemented human labor rather than replacing it. The most successful CEOs of the next decade won’t use AI just to cut entry-level headcount. They will use it to “supercharge” juniors—turning a 23-year-old with an AI co-pilot into the equivalent of a 30-year-old associate.
The “Canary” is singing. The question is: are you listening to the warning, or just enjoying the temporary silence in your payroll costs?
‘Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence’ by Erik Brynjolfsson, Bharat Chandar and Ruyu Chen – read here (PDF)
Take a look at previous posts in ‘The CEO’s guide to AI’ series:
About Nikolas Badminton
Nikolas Badminton is the Chief Futurist & Hope Engineer at futurist.com. He’s a world-renowned futurist keynote speaker, consultant, author, media producer, and executive advisor that has spoken to, and worked with, over 500 of the world’s most impactful organizations and governments.
Nikolas is an artificial intelligence expert and his 2026 keynote ‘The AI Leader: Create Incredible Productivity, Profit & Growth’ is the level up for the modern CEO and executive leader.
Please contact futurist speaker and consultant Nikolas Badminton to discuss your engagement.
