Three Perspectives for Leaders Setting 2027 Hiring Plans
Companies adopting artificial intelligence are expanding their workforces, but the growth sharply favors experienced employees over entry-level hires. A study published Sept. 21, 2026, by the Stanford Digital Economy Lab finds that firms using AI grow headcount by about 3.3% relative to comparable firms that have not adopted it while shifting its composition: senior employment rises 6.7% as junior employment falls 2.5%, a decline the authors find is not statistically significant.
Key Takeaways
- AI adopters are changing which workers they employ more than how many, with senior employment up 6.7% and the junior share down 1.9 percentage points across affiliates in 41 countries.
- Bharat Chandar and Bouke Klein Teeselink find the junior-share decline comes mainly from senior growth, while the 2.5% drop in junior headcount is not statistically significant.
- The junior share fell in 23 of 31 countries examined, with significant declines in seven, including the United States, Brazil and Saudi Arabia.
- Peter Cappelli argues that long-range AI displacement forecasts have missed and advises HR leaders to plan with BLS projections.
- Faisal Hoque argues that work redesign should start from what a role is for, citing reversals at Ford and Klarna.
Bharat Chandar: What the 41-Country Study Measured
Bharat Chandar of Stanford and Bouke Klein Teeselink of King’s College London analyzed 1.25 billion job postings and 154 million employment records across 41 countries. They identified AI adoption by searching job postings for generative AI terms such as ChatGPT, Claude, Gemini and prompt engineering, then used a large language model to confirm the role involved generative AI use. To separate the effect of adoption from other differences between firms, they compared employment trajectories of affiliates whose companies adopted AI to matched control affiliates in the same country and sector, using the AI adoption rate among other multinationals headquartered in the parent’s metro area as an instrument.
The research found that while overall employment at adopting firms grew modestly, the share of junior employees fell by 1.9 percentage points by March 2026. Senior employment shifted toward the most AI-exposed occupations by 0.9 percentage points, while junior employment shifted away from them by 0.6 percentage points, an estimate that is not statistically significant. Computer and mathematical occupations, the most AI-exposed group, saw their employment share grow by 0.8 percentage points while their junior share fell by 3.3 points.
The pattern appears across most of the world. The junior share of employment declined in 23 of 31 countries examined, with statistically significant drops in seven, including the United States, Brazil and Saudi Arabia. “AI is labor saving for junior workers and labor expanding for seniors in exposed occupations,” the authors wrote in “How Does AI Change Labor Demand? Evidence from 41 Countries.” Bloomberg Law covered the study the day it was released.
Peter Cappelli: Why AI Job Forecasts Keep Missing
In HRO Today (Sept. 24, 2026), Peter Cappelli, George W. Taylor Professor of Management at the Wharton School, argues that apocalyptic AI displacement forecasts have consistently missed and that conventional labor statistics remain the better planning input. In “At Least Two Cheers for the BLS, None for the AI Leaders,” Cappelli cites a study by Michael Handel at Northeastern University that followed up on AI predictions from the early 2010s. Handel compared reality to the U.S. Bureau of Labor Statistics Occupational Forecasts over the decade since. “The results show no support for any of the AI predictions,” Cappelli writes. “Overall, jobs didn’t decline, the jobs most at risk didn’t decline, and so forth.”
Cappelli identifies several reasons the forecasts fail. Tool developers are not experts on the organizational change processes needed to introduce tools, nor are they likely familiar with typical administrative work. Many projections were based on Moore’s Law, which stopped holding a decade or more ago. “There are lots of things in this world that we just cannot predict well,” he concludes. “There will always be people who tell you that they can predict. Just don’t believe them.”
Faisal Hoque: Why Work Redesign Starts With What a Role Is For
In Fast Company (Sept. 22, 2026), Faisal Hoque, founder of SHADOKA and NextChapter and an executive fellow at IMD, argues that the operational question is how to assess a role’s purpose before restructuring or eliminating it. In “Understand a Job Before You Eliminate It,” Hoque describes a five-step process for redesigning work in the age of AI, starting with what the organization needs the role for rather than what the person in it does.
Hoque cites Ford’s experience hiring about 350 veteran engineers, some of them former employees, to find the failure points its automated quality systems and AI inspection tools were missing. “The company had assumed that introducing artificial intelligence and feeding it design requirements would be enough to produce a high-quality product. It was not,” he writes. Klarna announced in February 2024 that its AI assistant was doing the work of 700 customer service agents; by May 2025 it was recruiting human agents again. A survey of 1,000 C-suite and senior leaders found that 39% had conducted layoffs as a result of AI, and 55% of those admitted to making wrong decisions about the redundancies.
Hoque’s process begins by describing the goals the organization needs the role for, then asks which goals need a person regardless of what the technology can do. For those that do not, it decomposes the work into tasks and tags each as automate, augment or not yet automatable. A fourth step chooses the lens for the redesign: cost, quality or human-centric. The final step recomposes what remains into whole roles. “Most AI redesigns start with a task list. That’s why they fail,” the article argues. “Start with what the role is for, and the picture changes completely.”
How Leadership Teams Can Plan the Junior Pipeline
Hiring plans for 2027 are being set now, and the entry-level rung is where AI shows up first in the numbers. A 6.7% senior increase alongside a 2.5% junior decline changes campus recruiting, apprenticeship budgets and the internal pipeline that produces senior staff five years out.
The question facing chief human resources officers is whether the company is cutting junior hiring on evidence or on expectation, and what it will cost to rebuild the rung if the forecasts are wrong. The specific decisions Hoque describes, whether to reserve a share of routine queries for new agents even though AI could handle them more cheaply, and whether to keep a visible route to a person because the customer is “a person before they are a ticket,” define what the redesigned workforce looks like and whose judgment it develops.
Bharat Chandar, Peter Cappelli and Faisal Hoque are each exclusively represented by Stellant Group for speaking engagements including keynotes, fireside chats and panel discussions with leadership teams thinking through AI’s effects on hiring, workforce planning and work redesign. For our consulting or advisory offerings, please reach out through stellantgroup.com for a custom proposal.
