Work & Distribution with AI
As machines take over more work, the question is how people secure income, purpose and participation, and who gains from the added productivity. Past technologies destroyed and created jobs, but transitions often took decades and hit regions unevenly. Social systems and taxes depend on wage labor, and whether AI acts differently this time is an open question. Entry-level hiring in exposed jobs is falling, while overall employment shows no clear AI-driven drop yet.
Open in the interactive tree →Historically, new technologies destroyed jobs and created new ones, but transitions often took decades and hit regions unevenly. Social systems and taxes depend on wage labor: contributions and income tax make up most revenue. Whether AI acts differently this time (broader, faster) is an open empirical question.
As of now
Entry-level hiring in exposed jobs is falling (Stanford, August 2026: employment of 22-25-year-olds in AI-exposed occupations is about 19% below that of less-exposed peers), while overall employment shows no clear AI-driven drop yet. The debate ranges from retraining and shorter working weeks to basic income, shared ownership and automation taxes; robust evidence for any single solution is missing as of October 2026.
What is missing
- Robust, near-real-time measurement of which tasks AI replaces or complements
- A financing model for social systems that depends less on wage labor
- Fast, effective retraining and job switching for adults
- Rules for spreading AI productivity gains broadly (ownership, taxes, participation)
- Evidence from well-evaluated pilot programs to build consensus on which models to scale
Becomes possible once solved
- Shorter working hours at maintained prosperity
- New occupations and wider participation
- Stable social systems with less paid employment
Open steps
- Real-time task-level measurement High AI leverageMeasure which tasks AI replaces versus complements, from usage logs, job postings and payroll data, and track it monthly.
- Causal effects on young workers Medium AI leverageSeparate AI's causal effect on entry-level hiring from interest rates, education mix and pre-existing trends.
- Financing social systems beyond wages Medium AI leverageModel tax and contribution systems that work if the wage share falls (capital, compute or automation taxes, ownership funds) and test their effects.
- Retraining that works for adults Medium AI leverageFind which retraining and job-switching programs raise adult employment, using well-evaluated pilots and personalized guidance.
Where today's AI could help
Medium AI leverage. AI helps measure task-level effects and test retraining; who owns and shares the gains remains a political decision.
- Measure in near real time which tasks AI automates or complements, from usage and job-posting data
- Simulate tax, ownership and basic-income designs in agent-based economies
- Personalized retraining and job matching for adult workers
- Evaluate pilot programs faster by analyzing outcome data
Shown so far
- In January 2026 Anthropic's Economic Index report tracked Claude use by task and found 52% of conversations augmenting work versus 45% automating it in November 2025, according to the company. source
- In May 2022 the AI Economist (Science Advances) used two-level reinforcement learning to find tax policies in simulated economies that improved the equality-productivity trade-off by at least 16% over baselines. source
Prerequisites
- Economics1776
- Social Insurance1883
- AI & the Labor Market2025-2026
Unlocks
- Participation Amid Automation2030s-2040s?