Workforce Rewired Daily Briefing | Sunday, May 17, 2026
New BLS occupational data confirms a second consecutive year of employment decline in AI-exposed roles. Separately, Stanford HAI and BambooHR released data on the specific population bearing the cost: workers under 25 and early-career hires in finance who are finding less and less work to do. GitLab named the organizational logic behind all three.
By the Numbers
0.2% decline in employment across 18 BLS-flagged AI-exposed occupations between May 2024 and May 2025, while overall U.S. employment grew 0.8% over the same period. (Bloomberg/BLS)
20% drop in employment for software developers ages 22 to 25 since 2022; CS undergraduate enrollment fell 11% between 2024 and 2025. (Stanford HAI 2026 AI Index)
1 in 3 junior hires in accounting and finance quit within their first year, in an environment where a 3-to-1 senior-to-junior hiring ratio leaves early-career workers with little substantive work to do. (BambooHR/Fortune)
40% of U.S. workers are in occupations with zero measured AI exposure; fewer than 10% are in roles with an AI exposure score of 0.4 or higher. (Federal Reserve Bank of New York)
Layoffs and Company Decisions
BLS Data Confirm a Second Year of Job Losses in AI-Exposed Roles
Annual occupational employment data from the Bureau of Labor Statistics, analyzed by Bloomberg, shows that 18 occupations the BLS has flagged as likely to be affected by AI shed a combined 0.2% of their workforce between May 2024 and May 2025. The broader labor market added 0.8% in headcount over the same period. Customer service representatives and certain secretary and sales roles led the losses. The 18 occupations together account for roughly 10 million workers. This is the second consecutive year of employment decline in these BLS-designated categories.
Source: Bloomberg / Bureau of Labor Statistics, May 15, 2026
Why it matters: Two consecutive years of decline in the same occupational categories removes the “cyclical noise” argument. HR leaders managing customer service, administrative support, and inside sales should treat this as structural pressure, not a temporary soft patch requiring routine management.
GitLab Removes Up to Three Management Layers in Shift to Agentic Model
GitLab announced a voluntary separation program alongside an organizational restructuring that removes up to three layers of management, reduces its country footprint by 30%, and reorganizes the company into approximately 60 autonomous R&D teams. The stated rationale is the “agentic era”: AI agents handling routine coordination and execution tasks, smaller human teams handling higher-judgment work, and management layers that previously existed to coordinate those tasks becoming structurally unnecessary. GitLab framed the move as a deliberate redesign rather than a cost-cutting layoff, though the savings will be reinvested into growth. Approximately 7% of roles are affected through the voluntary separation process.
Source: Bloomberg, May 11, 2026
Why it matters: GitLab is not describing a layoff, it is describing a career ladder that no longer needs certain rungs. Every organization should identify where its management layers exist primarily to coordinate work that AI is beginning to absorb, and decide whether that structure reflects deliberate design or inherited assumption.
New York Fed: Hiring Slowdown in AI-Exposed Jobs Predates ChatGPT
Researchers at the Federal Reserve Bank of New York published a job postings analysis using Anthropic’s task-level AI exposure framework to test whether firms cut hiring in AI-exposed occupations after ChatGPT launched in late 2022. Their finding: the divergence in postings between high-exposure and low-exposure occupations began before ChatGPT’s release, not after. There is also no evidence of a widening gap between junior and senior postings within high-exposure fields, which challenges the specific claim that AI is disproportionately eliminating entry-level roles in a measurable way. Fewer than 10% of workers and vacancies are in high-exposure occupations. New York Fed business surveys indicate firms plan to respond to AI primarily through retraining rather than headcount reduction.
Source: Federal Reserve Bank of New York, Liberty Street Economics, May 14, 2026
Why it matters: The Fed’s analysis complicates both the alarm and the dismissal. If the hiring slowdown in exposed roles predates generative AI, structural forces have been compounding longer than the current news cycle implies. The concrete takeaway: the window to build proactive retraining infrastructure may be shorter than the absence of regulatory pressure suggests.
Reskilling and Education
Stanford AI Index: Developer Employment Down 20% for Workers Under 25
Stanford HAI’s 2026 AI Index documents a 20% decline in employment for software developers ages 22 to 25 since 2022. CS undergraduate enrollment dropped 11% between 2024 and 2025. The students who chose coding as a durable career path are encountering a floor that has moved faster than their training anticipated. The Index also documents a widening gap between expert and public sentiment on AI’s labor market effects: 73% of AI researchers express optimism about long-run job creation, while 23% of the general public shares that view.
Source: Stanford HAI 2026 AI Index, May 2026
Why it matters: The 20% employment drop among young developers is a concrete data point on who absorbs early AI disruption. Universities building AI-integrated curricula and employers recruiting from CS programs should understand that enrollment decline and employment decline are now moving in the same direction simultaneously.
BambooHR: One in Three Junior Finance Hires Quit Within a Year as AI Absorbs Their Work
A BambooHR analysis of accounting and finance hiring, reported by Fortune, finds that one in three junior hires quit within their first year. The pattern behind the churn: AI tools are enabling senior workers to absorb tasks that previously required junior headcount, producing a 3-to-1 senior-to-junior hiring ratio in the sector. Early-career hires arrive expecting an on-ramp built on task volume and repetition. They find seniors who are not delegating the way seniors used to delegate, and roles that lack the substantive work needed to develop judgment and build skill. The pipeline is not just hiring less. It is retaining less of what it does hire.
Source: BambooHR / Fortune, May 12, 2026
Why it matters: High early-career churn in finance is a leading indicator for any function where AI is absorbing routine execution tasks. Organizations need an explicit answer for how skills develop when the repetition that used to build them is automated. “They’ll learn by doing” is not an answer if the doing has changed.
What Workforce Leaders Are Watching
The BLS data covers 10 million workers in AI-exposed occupations. What share of your organization’s headcount sits in those categories, and does your workforce plan treat them as under structural pressure or cyclical softness?
GitLab’s restructuring removed management layers whose primary function was coordinating work AI now handles. Which layers in your organization exist for the same reason, and is that a deliberate choice or an inherited structure you have not revisited?
If early-career churn is rising in your finance, operations, or administrative functions, the BambooHR pattern suggests the cause is not compensation or culture. It is an undefined role in an environment where the work that used to develop junior employees has been absorbed upstream. Are you tracking early-career attrition by function?
The NY Fed finds firms plan to retrain rather than cut. What governance structure makes that intention real? Who owns accountability when the retraining investment fails to produce the workforce the organization actually needs?
This briefing was prepared automatically by the Workforce Rewired research assistant. All stories include direct source links.