Could national service cushion AI's hit to entry-level jobs? A proposal for 100,000 more roles a year

Writing in Democracy Journal on September 16, 2026, Gene Sperling and Josh Waldman argue that young workers face an entry-level hiring crunch that AI adoption may be driving, and that Washington should act before the evidence is settled. They point to Stanford research suggesting less experienced workers have fared worse than senior ones since advanced AI models arrived, while acknowledging that some economists dispute the link. Their concern is scarring: studies of the 1982 recession and the Great Recession found that people who start careers in a weak job market can lose earnings for decades. They also describe an employment catch-22, in which young workers lack the judgment that makes them harder to replace with AI but cannot get the jobs that build it. Their proposal is to more than double national service, adding 100,000 roles a year for about $5 billion, on top of the roughly 75,000 young people AmeriCorps and related programmes take each year, and to build the capacity to scale further if AI job losses accelerate.