Workers ages 22 to 25 in the jobs most exposed to generative AI have seen employment drop 13% since 2022, according to a study by three Stanford University researchers. That was the year the first generative AI tools reached the public, and the decline falls hardest on the bottom rung of the career ladder.

The Stanford report states that AI adoption "is beginning to have a significant and disproportionate impact on entry-level workers in the American labor market." For HR leaders, the finding touches a core people-process decision: whether entry-level roles — the traditional proving ground for future leaders — still exist in sufficient numbers to sustain internal talent development.

Why entry-level roles matter beyond Gen Z hiring

Entry-level positions have long let early-career employees handle routine work: customer interactions, first drafts, research, data reporting. Through that work, they built both technical skills and human ones — problem-solving, collaboration, adaptability, empathy — and learned how the business operated and how decisions got made.

As AI absorbs those routine tasks, companies optimize short-term cost savings at the expense of long-term capability. The authors warn this threatens the default learning model of "hire smart people and have them learn on the job," because the remaining work requires time on task and experience to master. People do not develop judgment by watching videos, the piece argues; they learn by doing real work with real stakes.

What does the Taproot survey say about skills training?

A new report from the Taproot Foundation, based on a survey of people engaged in skills-based volunteering plus worker interviews, found that 82% of respondents believe human skills in the AI era are now as vital as technical skills. Yet only 16% — fewer than 1 in 5 — believe traditional training such as classroom learning and online courses can effectively develop creativity, critical thinking and real judgment.

That gap puts pressure on L&D design. If classroom formats cannot build judgment, employers need experiential structures that can.

How can employers restructure early-career roles?

The report and its authors propose three approaches:

• Human-AI pairing mentorships. Instead of eliminating all entry-level roles, pair new hires with experienced employees on shared projects. AI handles the data; the junior employee, under supervision, verifies outputs, weighs ethical implications and ensures accountability — then presents findings, building strategic communication skills.
• Skills-based promotion evidence. Move away from tenure-based credentials and make human skills visible through real-world project outcomes, whether internal or through pro bono engagements for nonprofits.
• Experiential sabbaticals. Pro bono work with local nonprofits offers complex, ambiguous, short-term leadership challenges outside the常规 training track.

The numbers behind the last approach: 73% of respondents found pro bono work effective at building the problem-solving and collaboration skills needed in an AI-driven economy. On the receiving end, 85% of nonprofits report that skills-based volunteers increase their long-term organizational capacity.

What's at stake for talent pipelines

The Stanford data suggests the squeeze on early-career hiring is already measurable, not hypothetical. Employers that let AI erase the routine work where judgment is formed will need a deliberate replacement — mentorship structures, project-based evidence and external experiential placements — or face a thinner internal bench for future leadership roles.