A surprising reversal: why “primary workers” are now absorbing the hiring shock
A recent Federal Reserve Bank of Richmond study highlights a labor-market dynamic that runs counter to decades of recession playbooks. So-called “primary workers”—roughly 55% of the U.S. labor force with stable employment histories—have seen their job-finding rate fall by 13 percentage points between November 2022 and September 2025. Meanwhile, “secondary workers”—about 14% of the workforce—experienced only a two-point decline.
That inversion matters because primary workers have historically been the most resilient cohort in downturns: they tend to have stronger networks, clearer career narratives, and credentials that travel well across employers. The new pattern suggests something more structural than a typical cyclical cooling. It implies that the labor market is not merely “slowing,” but re-sorting—with hiring pipelines, role definitions, and screening mechanisms shifting in ways that disproportionately penalize experienced workers seeking to move, re-enter, or reset.
Several signals reinforce the sense of a market that is tightening in nontraditional ways:
- Entry-level and routine white-collar roles—often the on-ramp for career transitions—are thinning out or being redesigned.
- Competition is intensifying for the remaining “safe” roles, pushing some seasoned candidates toward pay cuts and down-leveled titles.
- Even where certain sectors show pockets of growth (notably professional and business services, and information), net job losses in July underscore a market that remains fragile and selective.
The headline is not simply fewer jobs; it is fewer familiar pathways for workers who once relied on predictable ladders and transferable experience.
Generative AI exposure and the new mechanics of job-finding
The study’s most consequential finding is a strong negative correlation between job-finding rates and occupational exposure to generative AI, particularly since the public debut of ChatGPT in late 2022. In practical terms, the roles seeing the steepest hiring declines are those where large language models (LLMs) can replicate or accelerate meaningful portions of the workflow—especially routine, codifiable knowledge tasks.
This is not a simplistic “AI replaces workers” story. It is a story about how firms change hiring when AI changes throughput. Employers can maintain service levels, documentation output, analysis volume, and customer responsiveness with leaner teams, which alters the marginal value of adding a new hire. The result is a labor market where job creation can decelerate even as business activity remains stable.
Key labor-market mechanisms emerging from generative AI adoption include:
- Task automation and task compression: Roles in customer support, basic data entry, standard legal research, and template-based content work are increasingly supported by AI copilots, reducing the need for incremental headcount.
- A “productivity paradox” for hiring: AI can raise output per employee, but it can also raise expectations—new hires are increasingly expected to arrive with domain expertise plus AI fluency, rather than being trained into it.
- Credential inflation via tooling: Familiar tools (LLM interfaces, workflow automation, AI governance practices) become de facto requirements, even when job descriptions do not explicitly state them.
Importantly, this dynamic can hit primary workers harder than expected. Many experienced professionals are highly capable, but their advantage historically came from accumulated process knowledge—precisely the kind of knowledge that organizations now attempt to encode into playbooks, prompts, and automated workflows. When firms believe they can “bottle” routine expertise, they may reserve hiring for fewer, more specialized roles—leaving mid-career candidates competing for a narrower set of openings.
Corporate strategy under high rates: “do more with less” becomes an operating system
The macro backdrop—higher interest rates and tighter financial conditions—amplifies the labor effects of generative AI. When capital is expensive, executives face a dual mandate: control labor costs while still investing in digital transformation. Generative AI offers a compelling bridge: a technology spend that can substitute for incremental hiring, at least in the near term.
This is where the labor market begins to look polarized. The study’s implications align with a broader pattern of structural unemployment risk for mid-skill, routine white-collar work—roles that sit between high-end strategic knowledge work and in-person service work that is harder to automate.
From a business and technology perspective, three strategic shifts stand out:
- Budget reallocation from headcount to AI services: Firms increasingly treat AI as a scalable input—often purchased via platforms, vendors, and “AI as a service”—rather than expanding payroll.
- A bias toward proven AI-augmented experience: As roles evolve, employers may prioritize candidates who have already operated in AI-instrumented environments, raising barriers for workers transitioning across sectors.
- Wage pressure in the middle, premiums at the edge: Starting salaries and mid-tier compensation can soften where AI expands labor supply or reduces openings, even as firms pay up for niche expertise in AI engineering, governance, security, and domain-specific implementation.
This is also where macroeconomic feedback loops become plausible. If displaced or down-leveled workers experience sustained income compression, consumer demand can weaken—potentially reinforcing slower growth and slower hiring. The labor market then becomes both a symptom and a driver of economic cooling.
What a future-ready labor market will reward: hybrid capability, mobility, and human-AI workflow design
The most actionable takeaway is that generative AI is functioning less like a single technology wave and more like a re-architecture of work—changing how tasks are bundled, how teams are sized, and how careers progress. For employers, the strategic opportunity is to treat workforce redesign as a competitive advantage rather than a reactive cost exercise. For policymakers, the challenge is to prevent a widening gap between workers who can rapidly adapt and those stranded between old job definitions and new ones.
Forward-looking organizations are likely to differentiate themselves by building systems around human-AI collaboration, including:
- Cross-disciplinary talent pipelines that combine domain knowledge with AI literacy (rotational programs, apprenticeships, and micro-credential partnerships).
- Process reengineering that maps workflows for augmentation, reserving human attention for judgment-intensive work such as compliance, stakeholder management, and strategic decision-making.
- Continuous learning ecosystems that shift training from episodic courses to ongoing, subscription-like upskilling—keeping pace with fast-changing models, tooling, and governance norms.
- Internal mobility as a retention strategy, reducing the friction of career pivots and helping primary workers redeploy rather than exit.
The Richmond Fed’s findings land as a sober signal: the labor market is not only cooling—it is being re-indexed around AI-shaped productivity. The organizations that thrive will be those that can translate that productivity into durable value creation while keeping human capability moving up the stack, not out of the system.
By
By
By
By
By
By
