Is generative AI removing the bottom rung of careers?
Employment is rising as a keyword in the economics journals, and generative AI is growing there fast. Yet of nearly 1,200 recent papers on employment, wages and labour markets, only 14 carry a generative-AI keyword. Is AI taking the tasks that junior workers used to learn on, and what happens to their careers?
What prompted this
In the economics journals, "employment" had 16 papers in July–September 2026 against about 8 expected, "labor market tightness" 6 against 1, and "generative AI" 9 against 3 (see the economics dashboard). Since October 2025, 1,181 papers carry an employment, labour-market or wage keyword (918 in economics journals). Only 14 of them also carry a generative-AI, language-model or ChatGPT keyword, and "young workers" appears on 10.
Questions
- In occupations most exposed to generative AI, has hiring of new graduates fallen since 2023 relative to hiring of experienced workers?
- Do firms that adopt AI change their training and promotion ladders?
- Do workers who started their careers in exposed jobs earn less years later?
Data that could help
Online job postings (Lightcast), worker profiles (LinkedIn, Revelio), occupational task data (O*NET) to measure AI exposure, and payroll or administrative employment data.
Next step
Compare monthly entry-level and experienced job postings by occupational AI exposure from 2021 to 2026.