Goldman Sachs research published in early April 2026 finds that AI is eliminating a net 16,000 US jobs per month, with AI substitution removing approximately 25,000 positions monthly while augmentation creates only 9,000. The analysis draws on Bureau of Labor Statistics data cross-referenced with AI exposure indices.
Gen Z workers are disproportionately affected. They are concentrated in routine white-collar and administrative roles — data entry, customer service, legal support, billing — that generative AI automates most readily. Entry-level job postings have fallen roughly 35% since January 2023, according to Revelio Labs data cited alongside the Goldman analysis. Separately, McKinsey survey data indicates 51% of companies report that generative AI has reduced their demand for entry-level positions.
The wage and unemployment gaps between entry-level and experienced workers have widened most in sectors with high AI substitution risk. Goldman’s framing positions this as a structural shift in career ladders rather than a cyclical downturn: the traditional pathway from junior administrative role to mid-career professional is being compressed or eliminated by automation of the early rungs.
Relevance to the AI Gap thesis: This is direct evidence for the pre-redundancy training argument (dimension 8) and the case for worker rights in AI employment decisions (dimension 9). Workers at the bottom of the skills ladder are losing roles before training programmes reach them. The data supports the argument that employer AI training obligations need to be prospective, not reactive.