The current model of white-collar career progression involves taking an entry level job right out of school and doing codifiable tasks while slowly learning the tacit knowledge to become an experienced worker. That line from a Dallas Fed analysis captures the part of the AI labor shift that matters most right now: the strain is showing up less in broad layoffs than at the bottom rung of the ladder.
The clearest recent labor-market evidence does not show a simple machine-versus-human outcome. It shows a split. Jobs built around structured digital work are seeing the fastest AI penetration, especially roles involving drafting, summarizing, documentation, coding support, records handling, and scripted customer interactions. Anthropic’s task-based research placed programmers near the top, with customer service and data-entry roles close behind, using real workplace Claude usage rather than broad theoretical capability alone. That distinction is important because workplace adoption is not the same thing as technical possibility.
Even in heavily exposed fields, organizations are still deciding where AI belongs in real workflows, what can be trusted, and which tasks still need human oversight. Anthropic found that 33% of tasks in Computer & Math are currently covered, leaving a wide gap between what models may eventually handle and what employers are actually using them for today. Yale Budget Lab has also noted that exposure metrics increasingly diverge at the top end, meaning researchers broadly agree that software and knowledge jobs are highly exposed, while disagreeing on exactly how far that exposure goes in each occupation. The low end is much clearer: plumbers, cooks, bartenders, mechanics, and other hands-on roles remain far less exposed because physical environments, equipment handling, and embodied judgment still resist automation.
The more immediate disruption is turning up in hiring patterns. Anthropic’s data pointed to about a 14% decline in entry into highly exposed jobs for workers ages 22 to 25 relative to 2022. Dallas Fed researchers describe the same pressure in broader terms: young workers are facing a tougher path into AI-exposed sectors even while unemployment has not surged for established employees. The mechanism is straightforward. Entry-level work often leans on codified knowledge that can be written down, standardized, and increasingly handled by AI systems, while experienced workers contribute tacit judgment that is harder to replicate. That helps explain why the wage story looks less bleak than the hiring story.
In AI-exposed sectors, employment has softened in some categories, but pay has not broadly collapsed. Dallas Fed data showed computer systems design wages rose 16.7% since fall 2022, well above the national average in the same period. Separate research cited by Fortune said highly exposed occupations have still delivered real wage increases and job growth. Microsoft researchers added a similar caution against easy conclusions: “Our data do not indicate that AI is performing all of the work activities of any one occupation.”
What emerges is not a clean replacement map but a redesign map. Knowledge work remains highly exposed, especially where tasks are text-heavy and digital, yet the people with experience are often being complemented rather than displaced. The pressure is building first where careers begin, not where they peak.
