RAND: Social & Economic Policy in the Age of AI

Executive Summary

RAND RRA3888-2. Examines AI adoption in four sectors: financial services, healthcare, climate/energy, transportation. Explores the dual macroeconomic effects of AI: enhanced productivity gains vs. substantial labor displacement. By 2035, moderate AI-driven productivity gains could boost real per-capita GDP by nearly $7,000.

General-purpose AI systems capable of performing across diverse cognitive tasks are poised to reshape labour markets, wealth distribution, and social safety net requirements in ways that existing policy frameworks cannot adequately address. This research examines the macroeconomic and social policy implications of widespread AI adoption, arguing that incremental adjustments to current regulatory and welfare structures will prove insufficient. The study proposes a comprehensive policy rethinking spanning workforce transition support, education system transformation, progressive taxation of AI-driven productivity gains, and updated social insurance mechanisms designed for an economy where human labour's share of value creation may substantially diminish. Drawing on economic modelling and comparative policy analysis across advanced and developing economies, the research provides evidence-based recommendations for policymakers navigating the transition to an AI-augmented economic paradigm.

Key Findings

34%

General-purpose AI amplified labour market polarisation with routine cognitive roles facing disproportionate displacement pressure

Of occupations classified as routine cognitive experienced measurable task automation within eighteen months of enterprise GenAI deployment, concentrated in administrative and clerical functions.

$1.8T

Universal basic income simulations incorporating AI productivity gains showed fiscal viability under moderate adoption scenarios

Projected annual productivity surplus available for redistribution in advanced economies under median AI adoption projections, potentially funding universal basic income programmes.

7.3 yrs

Education system reform timelines lagged AI capability advancement by a widening margin, creating persistent skill mismatches

Average lag between emergence of new AI-adjacent occupational requirements and corresponding curriculum availability in national education systems across OECD member countries.

62%

Progressive taxation of AI-derived economic rents emerged as the most politically viable funding mechanism for transition support

Public support for targeted taxation of corporate AI productivity gains when revenues were earmarked for worker retraining and transition assistance programmes.

Abstract

About This Research

Publisher: RAND Corporation Year: 2025 Type: Applied Research

Source: Rethinking Social and Economic Policy in the Age of General-Purpose AI

Relevance

Industries: Financial Services, Government, Healthcare Pillars: AI Readiness & Strategy Use Cases: Cybersecurity & Threat Detection

Labour Market Displacement and Transition

The economic modelling projects significant occupation-level disruption concentrated in routine cognitive tasks including data entry, basic analysis, content creation, and customer service. However, the research challenges simplistic narratives of mass unemployment, demonstrating that AI adoption simultaneously creates demand for new roles in AI system oversight, human-AI collaboration design, and sectors where human qualities such as empathy, physical dexterity, and contextual judgement remain economically valuable. The critical policy challenge lies in managing the transition velocity—ensuring that displaced workers can access retraining and redeployment support before economic hardship becomes entrenched.

Education System Transformation

Current education systems optimised for knowledge transmission face obsolescence as AI systems increasingly outperform humans at information retrieval and synthesis. The research advocates for fundamental curriculum restructuring that prioritises metacognitive skills, creative problem-solving, ethical reasoning, and human-AI collaboration competencies. Lifelong learning infrastructure must replace the front-loaded education model, providing accessible reskilling opportunities throughout careers that may span multiple AI-driven occupational transitions.

Progressive Taxation and Redistribution

If AI substantially increases productivity while reducing labour demand, existing tax systems predicated on income and payroll taxation will face eroding revenue bases precisely when demand for social support escalates. The research examines alternative taxation mechanisms including AI productivity levies, automation taxes calibrated to displacement impact, and broadened capital gains taxation designed to ensure that AI-driven wealth creation funds adequate transition support and social infrastructure.

Key Statistics

34%
of routine cognitive occupations saw measurable task automation
Rethinking Social and Economic Policy in the Age of General-Purpose AI

$1.8T
annual productivity surplus potentially available for redistribution
Rethinking Social and Economic Policy in the Age of General-Purpose AI

7.3 yrs
lag between new skill demands and curriculum availability
Rethinking Social and Economic Policy in the Age of General-Purpose AI

62%
public support for taxing AI gains to fund retraining
Rethinking Social and Economic Policy in the Age of General-Purpose AI

Common Questions

The research examines several complementary mechanisms including graduated automation levies that scale with the degree of human labour displacement attributable to AI systems, broadened capital gains taxation that captures a larger share of AI-driven productivity increases, and digital services taxes applied to revenue generated through substantially automated business processes. The modelling suggests that a combination of these instruments can maintain social safety net funding at current levels even under aggressive AI adoption scenarios, though the optimal policy mix varies significantly across national economic contexts.

The computational modelling projects net employment effects ranging from modest decline to slight increase depending on the adoption scenario, national economic structure, and policy response adequacy. Crucially, the research emphasises that aggregate employment figures mask substantial compositional shifts, with routine cognitive occupations experiencing significant contraction while demand for AI oversight, creative, and interpersonal roles expands. The distributional impact is highly uneven, with lower-skilled workers in routine roles bearing disproportionate transition costs unless proactive policy intervention provides accessible reskilling pathways and interim income support.

Rethinking Social and Economic Policy in the Age of General-Purpose AI