The problem HR is facing
One of the most discussed findings on HR sites in 2026 comes from Stanford's Digital Economy Lab. Erik Brynjolfsson, Bharat Chandar and Ruyu Chen analysed payroll data from ADP covering millions of US workers. In their August 2026 update:
- Employment of 22 to 25 year olds in AI exposed jobs is 19% below where it would be if it had kept pace with less exposed peers.
- Experienced workers show no comparable gap.
- The decline sits in jobs where AI mainly replaces tasks. Where AI mainly supports workers, employment is flat or rising.
- The authors stress they find no evidence of economy wide job loss. The pressure is at the entry level.
It looks efficient today. The bill arrives in five to ten years, when the seniors retire and nobody learned the job.
Your talent pipeline in 8 years
A simple model of one team. Each year you hire juniors, some become seniors after a few years, and some seniors leave. All numbers are your own assumptions.
A way forward
The Stanford finding contains its own hint: where AI supports people instead of replacing tasks, employment holds up. So the question for HR is not "do we still need juniors?" but "what is a junior job when AI does the routine part?"
- Redesign junior roles around judgement. Juniors check, question and improve AI output with a senior, instead of doing the routine work AI now does.
- Make mentoring part of the senior job. If seniors are measured only on output, they will hand the routine work to AI and nobody learns. Give mentoring time and credit.
- Plan the bench, not just this year's budget. Run the pipeline numbers for five and ten years before cutting junior hiring.
- Keep apprenticeships and internships. They are the cheapest way to keep the door open while roles change.
What this means for HR
- This is strategic workforce planning. HR is the function that sees the five year bench. Bring the pipeline numbers to the leadership table.
- Watch for a hidden age effect. Cutting entry level hiring mostly affects young people. Check how your hiring decisions affect different groups.
- Learning paths need a rewrite. If the first two years no longer teach the basics through routine work, onboarding and training have to teach them on purpose.
Honest limits
The Stanford figures are US payroll data and descriptive, not proof that AI caused the gap. Other analyses find it hard to separate AI from other reasons for weak entry level hiring, such as the general economy. The pipeline simulator is a simple illustration with assumptions you set yourself.
Sources
- Brynjolfsson, E., Chandar, B., & Chen, R. (2025, revised August 2026). Canaries in the coal mine? Six facts about the recent employment effects of artificial intelligence. Stanford Digital Economy Lab. source