The Productivity-Pay Gap in the Age of AI: Divergence, Dynamics, and Corrective Initiatives

I. Executive Summary

The persistent and growing divergence between productivity growth and the compensation of typical workers, commonly referred to as the productivity-pay gap, stands as a central challenge to equitable economic growth and shared prosperity.¹ This report examines the potential impact of Artificial Intelligence (AI) on this gap, with a particular focus on the analytical framework provided by the Economic Policy Institute (EPI). The EPI posits that this divergence, particularly evident since the late 1970s in the United States, is not an inevitable consequence of economic forces but rather the result of specific policy choices that have eroded the bargaining power of labor.²

The emergence of AI introduces a powerful new dynamic. AI technologies possess a dual capacity: they could exacerbate the existing productivity-pay gap by further concentrating economic gains and disempowering segments of the workforce, or, under appropriate policy and institutional frameworks, they could contribute to narrowing the gap by augmenting worker capabilities, democratizing skills, and fostering new avenues for value creation.³ The ultimate trajectory is not technologically predetermined but will be shaped by societal choices and interventions.

Currently, a range of initiatives are underway to navigate AI's impact. These include governmental policy and regulatory proposals, extensive workforce development programs spearheaded by public and private entities, corporate reskilling efforts, non-profit interventions aimed at equitable access and education, and strategic adaptations by labor unions seeking to protect worker interests in an AI-driven economy.⁵

This report concludes that AI's influence on the productivity-pay gap is contingent upon the policy environment in which it is deployed. Proactive, worker-centric policies, consistent with the EPI's recommendations for rebalancing labor market power, are crucial if AI-driven productivity gains are to translate into broadly shared prosperity.¹⁹ A coordinated strategy is essential to ensure that technological advancement aligns with the goals of reducing economic inequality and enhancing the well-being of the typical worker. The overarching consideration is that technology, including AI, operates within a socio-economic and policy framework; its effects are largely determined by this framework, not by its inherent characteristics alone.

II. The Productivity-Pay Gap: A Deep Dive into EPI's Analysis

A. Defining the Productivity-Pay Gap

The term "productivity-pay gap" specifically refers to the observed divergence between the growth of labor productivity and the growth of compensation for the typical worker.¹ Labor productivity is a measure of the economy-wide income generated per hour of work, reflecting the potential for rising living standards. "Pay" or "compensation" in this context, as defined by the Economic Policy Institute (EPI), encompasses the average wages and benefits of production and nonsupervisory workers. This group constitutes approximately 80% of the private-sector U.S. workforce, thereby representing the economic experience of the vast majority of workers, distinct from highly paid managerial and executive staff.²

It is important to distinguish this definition from other uses of the term "productivity gap." For instance, the term can also describe a sustained difference in measured output per worker (or GDP per person employed) between countries.²² Alternatively, it can refer to the "production gap," which is the difference between an economy's or organization's potential output and its actual output, highlighting inefficiencies.²³ While these broader concepts are relevant to overall economic performance, this report adheres to the EPI's specific definition, which focuses on the distributional question of how the fruits of increased efficiency are shared with the workforce.

B. Historical Divergence: Trends in Productivity vs. Compensation

The historical data presented by the EPI reveals a stark shift in the relationship between productivity and typical worker pay in the United States. For several decades following World War II, from 1948 until the late 1970s, the economic gains from rising productivity were broadly shared. During this period, the compensation of production and nonsupervisory workers grew in close alignment with net productivity growth.² Specifically, from 1948 to 1979, net productivity (economy-wide productivity net of depreciation) increased by 112.5%, while the real hourly compensation of a typical worker grew by a substantial 90.2%.²⁵ This near lockstep growth indicated that the economic system was effectively translating efficiency gains into improved living standards for the majority of the workforce.

However, a significant decoupling began in the late 1970s. According to updated EPI analysis covering the period from the fourth quarter of 1979 (1979q4) to the first quarter of 2025 (2025q1), net productivity grew by 86.0%. In stark contrast, the real hourly compensation for the typical worker increased by only 32.0% over this same extended period.² This means that productivity grew 2.7 times as much as pay for the median worker. An earlier analysis covering 1979 to 2019 showed an even wider divergence, with productivity up 85.1% and typical worker compensation up only 13.2%, a 3.5-fold difference.²⁵ This dramatic divergence underscores a fundamental shift in how economic gains have been distributed.

This period of decoupling also occurred within a broader historical context of fluctuating productivity growth rates. The post-World War II era saw significant productivity gains, which decelerated after 1973 but remained above earlier historical rates for many industrialized nations.²⁶ The United States experienced a productivity surge in the late 19th century and again in the mid-20th century, with real GDP per hour worked averaging almost 2.5% growth from 1913 to 1950.²⁶ The divergence identified by EPI thus represents a specific distributional phenomenon within these larger macroeconomic trends.

To illustrate this divergence clearly:

Table 1: Historical Divergence of Productivity and Typical Worker Compensation (Based on EPI Data)

Period	Net Productivity Growth (%)	Typical Worker Hourly Compensation Growth (%)	Ratio of Productivity Growth to Pay Growth
1948–1979	112.5	90.2	1.25 : 1
1979q4–2025q1	86.0	32.0	2.69 : 1

Source(s): ²

This table starkly visualizes the breakdown of the post-war linkage between productivity and pay, establishing the empirical basis for the concerns addressed in this report.

C. EPI's Framework: Policy Choices and the Widening Gap

The EPI's central argument is that the widening productivity-pay gap is not an inevitable outcome of abstract economic forces like globalization or automation alone. Instead, it is primarily the result of a series of deliberate policy shifts enacted since the late 1970s that have systematically eroded the economic leverage and bargaining power of typical workers.²

During the period of broadly shared prosperity (roughly 1948 to the late 1970s), a different set of policy priorities prevailed. These "policy bulwarks," as termed by the EPI, fostered an environment where workers could claim a fair share of economic growth. Key elements included:

• Macroeconomic policy targeting "high-pressure" labor markets: This meant aiming for sustained low unemployment, which enhances worker bargaining power.²
• A rising federal minimum wage: Regular and significant increases in the minimum wage helped to lift the floor for low-wage workers and had spillover effects on wages above the minimum.²
• Protection of unionization rights: Labor laws and their enforcement actively safeguarded workers' rights to organize and bargain collectively, allowing unions to secure better wages and benefits.²
• High top tax rates: More progressive taxation, including high marginal tax rates on top incomes, may have tempered excessive executive compensation and provided revenue for public investments.²
• Regulations against anti-worker practices: Various regulations constrained corporate power and practices that could harm worker interests, including in areas like anti-trust and financial markets.²

Beginning in the late 1970s, these policy supports were systematically dismantled or weakened. The EPI identifies several key shifts:

• Tolerance of excess unemployment: Macroeconomic policy became more focused on controlling inflation, often at the expense of full employment, leading to periods of higher unemployment that weakened worker leverage.²
• Stagnation of the minimum wage: Increases in the federal minimum wage became smaller and less frequent, causing its real value to erode significantly over time.²
• Failure of labor law: Labor law did not adapt to increasing employer hostility towards unions, making it more difficult for workers to organize and bargain collectively. Union density declined sharply during this period.²
• Lower tax rates on top incomes: Significant reductions in tax rates for high earners and corporations altered incentives and income distribution.²
• Anti-worker deregulation: Deregulatory initiatives in sectors such as trucking and airlines, alongside changes in anti-trust policy and financial regulation, often shifted power towards employers and capital owners.²

The cumulative effect of these policy changes, according to the EPI, was the suppression of wage growth for the vast majority of the workforce (the bottom 80%). The economic gains generated by ongoing productivity growth were not lost from the economy; instead, they were redirected. The EPI identifies two primary destinations for this "missing" pay:

1. Higher salaries for highly paid corporate executives and professional employees: This contributed to a significant increase in wage inequality within the labor market.²
2. Increased corporate profits and returns to capital owners: A larger share of national income flowed to shareholders and other owners of capital, rather than to labor compensation.²

These two factors—the concentration of labor income at the very top and the shift of income from labor to capital—are identified by the EPI as the primary drivers of overall economic inequality since the late 1970s.²

Table 2: EPI's Analysis of Policy Shifts and their Impact on the Productivity-Pay Gap

Policy Area	Policy Environment (1948-Late 1970s)	Policy Environment (Post-Late 1970s)
Labor Market Conditions	Targeting low unemployment (high-pressure labor markets)	Tolerance of higher unemployment to control inflation
Minimum Wage	Rapid and regular increases in federal minimum wage	Infrequent and smaller increases; erosion of real value
Unionization	Active protection of unionization rights; robust union density	Failure of labor law to counter employer hostility; declining union density
Tax Rates	High top marginal income tax rates	Significant reductions in top tax rates
Regulation	Regulations against anti-trust and other anti-worker corporate practices	Deregulation in various sectors; weakened anti-trust enforcement

Source(s): ²

The framework established by the EPI implies that any new technological advancements, such as AI, will inevitably interact with a labor market already profoundly shaped by these decades of policy choices. These policies have tilted bargaining power away from typical workers. Consequently, the impact of AI on wages and the productivity-pay gap will be filtered through this pre-existing imbalance. If AI automates tasks or alters skill demands, its effect on worker compensation will largely depend on the ability of workers to bargain for a share of any new productivity gains. Given the compromised state of worker power described by the EPI, there is a significant risk that, without countervailing policy interventions, AI could further entrench or even widen the existing gap. This is not necessarily due to the inherent nature of AI technology itself, but rather because of the labor market context into which it is being introduced. This path dependency is a critical consideration for understanding AI's potential future impact.

Furthermore, the EPI's identification of where the "missing" compensation has gone—to higher profits and the salaries of top earners—highlights that the productivity-pay gap is fundamentally an issue of income distribution, not a mysterious evaporation of economic value. This has direct implications for how AI-driven productivity gains might be distributed. AI is projected to create substantial economic value and productivity enhancements. The crucial question, therefore, is who will capture this newly generated value. If the policy environment and power dynamics that led to the historical productivity-pay gap persist, it is highly probable that AI-generated gains will also disproportionately flow to capital owners and the highest-paid employees. This scenario would likely exacerbate the productivity-pay gap and overall economic inequality, unless specific, robust measures are implemented to ensure a broader and more equitable distribution of these benefits.

III. Artificial Intelligence: A New Frontier for Productivity and Labor

A. The Promise and Peril of AI for Economic Output

Artificial Intelligence, particularly the rapid advancements in generative AI, is widely anticipated to be a transformative force for global economic output. Projections from entities like McKinsey & Company suggest that AI could contribute as much as $13 trillion to the global economy, signaling a new wave of technological potential.²⁷ This optimism is rooted in AI's capacity to automate a wide array of tasks, enhance complex decision-making processes through advanced data analysis, and foster innovation that could lead to the creation of entirely new job categories and industries.²¹ AI's applications span diverse sectors, promising efficiency gains in areas from healthcare and finance to manufacturing and customer service.²⁷

However, this technological promise is accompanied by significant concerns regarding its impact on labor markets and economic equality. A primary apprehension is the potential for widespread job displacement as AI systems become capable of performing tasks previously undertaken by humans.⁴ The International Monetary Fund (IMF) estimates that almost 40% of global employment is exposed to AI, with this figure rising to about 60% in advanced economies due to the prevalence of cognitive-task-oriented roles that AI can increasingly address.⁴ Beyond displacement, there are fears of skill polarization, where demand for very high-level AI-complementary skills and very low-level service skills might increase, hollowing out opportunities for mid-skill workers. This could deepen income inequality, a risk flagged by institutions like the World Economic Forum.⁴

B. Solow's Paradox Revisited: AI Implementation Lags and Early Productivity Signals

Despite the significant advancements and increasing adoption of AI technologies, their impact on aggregate productivity statistics has, to date, been less clear than many might expect. This situation echoes the "Solow paradox," famously articulated by economist Robert Solow in 1987 with respect to computers: "You can see the computer age everywhere but in the productivity statistics".²⁹ Current observations suggest we may be experiencing a similar phenomenon with AI.²⁹

Several explanations have been proposed for this apparent lag between technological capability and measured economic impact. These include the possibility of initial over-optimism ("false hopes"), difficulties in accurately measuring productivity gains from new technologies ("mismeasurement"), the chance that gains are being captured by a narrow segment, thus not appearing in broader averages ("redistribution"), and, perhaps most significantly, "implementation lags".³⁰ General-purpose technologies (GPTs) like AI, electricity, or the personal computer historically require substantial time to diffuse widely and for their full economic effects to be realized. This is because their benefits often depend on the development and adoption of complementary innovations, significant organizational restructuring, the acquisition of new skills by the workforce, and overcoming various adjustment costs.²⁹

Early evidence on AI's productivity impact is mixed and often context-specific. Some firm-level studies and experiments have reported substantial productivity increases. For instance, research has shown a 14% productivity boost for customer care agents using generative AI ²⁹, and another study reported a remarkable 66% increase in employee productivity through the adoption of generative AI tools in specific tasks.²⁷ However, broader economic studies have found minimal impact on aggregate earnings or hours worked so far.³² One analysis of Danish workers using AI chatbots found only minor time savings (2.8%) and that these savings rarely translated into higher earnings, with only a small fraction (3-7%) of productivity benefits being passed on to workers in the form of increased pay.³² This discrepancy between micro-level potential and macro-level realization is characteristic of the early phases of GPT adoption. Economic literature suggests that GPTs can even initially lead to a temporary fall in productivity as organizations grapple with the necessary changes in production methods, company structure, and human capital investment, often following a "J-curve" pattern where productivity growth first declines before accelerating.²⁹

C. Historical Parallels: Lessons from Previous Technological Revolutions (e.g., Internet, PCs)

Drawing parallels with previous technological revolutions can offer valuable insights into the potential trajectory of AI's economic impact. Technologies like electricity, the personal computer (PC), and the internet are considered general-purpose technologies that, like AI, fundamentally altered economic processes and societal structures.²⁹ These earlier GPTs also exhibited delayed productivity impacts. For example, electricity was introduced in the late 19th century, but its significant effects on productivity were not fully realized until after World War I, decades later.²⁹ Similarly, the internet and information and communication technologies (ICTs), while transformative, took time to translate into broad productivity gains, with a notable boost observed between 1995 and 2000.³⁵

Previous technological waves often reshaped labor markets by altering the demand for different types of skills, frequently favoring higher-skilled workers—a phenomenon known as skill-biased technical change.³ Studies found that workers who used computers on the job earned wages 10 to 15 percent higher than those who did not, even after controlling for other factors, suggesting a wage premium for computer skills.³⁷ Access to a PC and the internet at home was also correlated with higher wages, potentially due to better job search capabilities and skill development.³⁸

However, there are arguments that AI might differ from these historical precedents in significant ways. One key distinction is AI's potential to automate or augment tasks previously considered non-routine and requiring high levels of cognitive skill, which were largely insulated from earlier waves of automation.³ This could mean that AI's impact on the labor market, particularly for high-skilled, high-wage earners, may follow a different pattern than that observed with PCs or the internet.

The "implementation lag" observed with AI is not merely a technical hurdle; it represents a profound organizational and human capital challenge. Historical parallels suggest that realizing the full productivity potential of a GPT like AI necessitates substantial co-investments in redesigning business processes, developing new workforce skills, and adapting organizational structures to leverage the new technology effectively.²⁹ Simply deploying AI tools without these complementary changes is unlikely to yield significant or widespread benefits. This underscores the importance of policies and initiatives that support not just AI technology development, but also the broader ecosystem of organizational adaptation and workforce preparedness.

The current discourse surrounding AI, characterized by a wide spectrum of predictions from immense optimism about productivity and job creation ²⁷ to more cautious or even pessimistic views regarding its immediate economic impact and potential for job displacement ³², mirrors the uncertainty that accompanied previous major technological shifts. The "clash of expectations and statistics" noted for AI ³⁰ is reminiscent of the "exaggeration and unrealistic business models" that characterized the early internet era.³⁵ This historical pattern suggests that we are currently in a period of high uncertainty typical of the early stages of a GPT's diffusion. Consequently, both policy and business strategies need to be adaptive and resilient, avoiding over-reliance on definitive forecasts. Instead, the focus should be on building mechanisms that allow for flexibility and ensure that benefits, when they materialize, are shared equitably, regardless of the precise technological and economic trajectory AI ultimately follows.

IV. AI's Potential Impact on the Productivity-Pay Gap

Artificial Intelligence stands at a crossroads in its potential influence on the long-standing productivity-pay gap. Its capabilities could drive forces that either narrow this divergence by empowering a broader range of workers or exacerbate it by concentrating gains among a select few and further marginalizing others.

A. Mechanisms for Narrowing the Gap

Several characteristics of AI technology suggest potential pathways to a more equitable distribution of productivity gains:

• Augmenting Low- and Mid-Skill Workers: AI tools, particularly generative AI, can act as powerful assistants, enhancing the productivity and capabilities of workers who may have less experience or fewer specialized skills. By providing real-time information, automating routine aspects of complex tasks, and offering decision support, AI could "level the playing field".³ For example, a field study involving software developers found that AI assistants enabled the least-experienced and lower-paid individuals to achieve the most significant relative productivity gains, suggesting a potential compression of skill-based wage differentials from the bottom up.²⁸
• Democratizing Access to Skills and Tools: AI has the potential to lower barriers to acquiring and utilizing sophisticated skills. Complex analytical tasks, content creation, or even coding, which previously required extensive training, can be made more accessible through AI-powered tools.³ This democratization could empower a wider segment of the workforce to perform higher-value tasks, potentially increasing their earnings.
• Task Automation for High-Skill, High-Wage Earners: Unlike many previous waves of automation that primarily targeted routine, often lower-to-middle-skill tasks, AI is increasingly capable of performing complex, non-routine cognitive tasks. These include activities traditionally undertaken by highly educated and highly paid professionals, such as diagnosing diseases, conducting legal research, developing financial models, or writing sophisticated computer code.³ If AI automates portions of these high-end jobs, it could reduce the demand for some highly specialized human labor, potentially exerting downward pressure on wages at the upper end of the distribution and thereby narrowing the overall wage gap.

B. Factors Potentially Exacerbating the Gap

Conversely, several factors suggest that AI could reinforce or even widen the existing productivity-pay gap:

• Skill-Biased Technical Change and High-Skill Complementation: AI may disproportionately benefit workers who already possess high levels of skill and education, particularly those who can effectively use AI to augment their existing capabilities. This "skill-biased technical change" could lead to further increases in productivity and wages for top earners, while those unable to adapt or whose skills are substituted by AI fall further behind, leading to increased wage polarization.⁴ Studies suggest higher adoption rates of AI among high-wage, highly educated workers, which could initially widen labor market inequality.³¹
• Increased Returns to Capital Owners: As firms adopt AI to enhance productivity and reduce labor costs, a significant portion of the economic gains may accrue to the owners of capital (e.g., shareholders) rather than to labor in the form of higher wages.² This aligns with the EPI's historical analysis, which found that a substantial share of the productivity gains since the 1970s has flowed to profits rather than to typical worker compensation.² If AI accelerates this trend, the gap between productivity growth and pay growth for the majority will likely widen.
• Job Displacement and Wage Stagnation for Certain Groups: For tasks that AI can fully automate, there is a risk of direct job displacement or a reduction in labor demand, leading to unemployment or downward pressure on wages for affected workers.⁴ Estimates, such as Goldman Sachs' projection that generative AI could expose 300 million full-time jobs worldwide to automation, highlight the potential scale of this disruption.²⁷ Certain demographics, such as older workers, may be particularly vulnerable due to challenges in re-skilling or transitioning to new roles.²¹
• Uneven Global Impact and Digital Divide: The capacity to develop, deploy, and benefit from AI is not evenly distributed globally. Richer countries and firms with greater resources are generally better positioned to harness AI's advantages, potentially widening economic disparities between and within nations.²¹ This can indirectly affect the domestic productivity-pay gap through various channels, including shifts in global competitiveness, patterns of international trade, and decisions regarding offshoring or reshoring of economic activities.

The following table summarizes these contrasting potentials:

Table 3: AI's Dichotomous Potential: Shrinking vs. Widening the Productivity-Pay Gap

Factor	Mechanisms for Narrowing the Gap	Mechanisms for Widening the Gap
Impact on Skill Premium	AI augments less-skilled workers, democratizes skills, potentially reducing demand for some high-skill tasks.3	AI disproportionately complements high-skilled workers, increasing their wages and productivity further (skill-biased technical change).4
Returns to Capital vs. Labor	If AI empowers workers to create more value and bargain effectively, a greater share of gains could go to labor.	Productivity gains from AI boost capital returns, shifting income from labor to capital owners, especially if worker bargaining power is weak.2
Job Creation/Displacement	AI creates new roles focused on AI development, management, and human-AI collaboration; augments existing roles making them more valuable.21	AI automates tasks leading to job displacement or reduced hiring for certain roles, potentially depressing wages for affected workers.4
Access to Technology	AI tools become widely accessible and affordable, enabling broader participation in value creation.	Unequal access to AI technology, data, and AI-related education/skills concentrates benefits among early adopters or privileged groups.21

Source(s): ², ³, ⁴, ²¹

The distinction between AI's role in "augmentation" versus "automation" is particularly salient.⁴¹ If AI primarily serves to augment human capabilities across various skill levels, it is more likely to foster shared productivity gains and potentially narrow the pay gap, provided workers can bargain for these gains. Conversely, if AI's primary function becomes direct automation and displacement of human labor without commensurate creation of new, well-compensated roles or strong redistributive policies, the gap is likely to widen. Initial analyses, such as the Anthropic Economic Index, suggest a slight lean towards augmentation (57%) over automation (43%) in current AI usage, particularly in fields like software development and technical writing.⁴¹ While this offers a tentatively positive signal, the rapid evolution of AI capabilities means this balance could shift. This underscores the importance of policies and development practices that intentionally steer AI towards augmentation and human-AI collaboration.

Furthermore, the impact of AI on the productivity-pay gap will be significantly shaped by pre-existing inequalities and differential access to the technology and the skills required to use it effectively. Groups that are already marginalized, or those with limited access to education and reskilling opportunities, may find themselves disproportionately disadvantaged by AI-driven economic changes, even if AI generates overall productivity benefits.⁴ AI could lead to polarization within income brackets, where individuals and firms capable of harnessing AI see their prospects improve, while others fall further behind. This risk is compounded by the observation that wealthier nations and entities are often better equipped to capitalize on AI, potentially reinforcing existing disparities.⁴⁰ If access to AI tools, the training to use them, and the jobs that complement AI are unevenly distributed, then even substantial AI-driven productivity growth may not translate into better outcomes for typical workers. Instead, it could worsen their relative economic position, thereby further widening the productivity-pay gap. This potential outcome reinforces the EPI's central argument about the critical importance of robust mechanisms for equitable distribution and the empowerment of workers.

Works Cited (Click to Expand/Collapse)

1. clockify.me, accessed June 8, 2025, https://clockify.me/productivity-pay-gap...↩
2. The Productivity-Pay Gap | Economic Policy Institute, accessed June 8, 2025, https://www.epi.org/productivity-pay-gap/↩
3. Could AI Level the Playing Field of Earnings Inequality?, accessed June 8, 2025, https://ifr.org/post/could-ai-level-the-playing-field-of-earnings-inequality↩
4. AI Will Transform the Global Economy. Let's Make Sure It Benefits Humanity., accessed June 8, 2025, https://www.imf.org/en/Blogs/Articles/2024/01/14/ai-will-transform-the-global-ec...↩
5. Governments Prepare for AI Revolution to Boost Workforce Productivity - CoinCentral, accessed June 8, 2025, https://coincentral.com/governments-prepare-for-ai-revolution-to-boost-workfo...↩
6. Global study: AI-driven government productivity efforts can't underestimate culture - PR Newswire, accessed June 8, 2025, https://www.prnewswire.com/news-releases/global-study-ai-driven-government...↩
7. The reskilling roadmap: Navigating evolving human, AI roles in the workplace, accessed June 8, 2025, https://www.chieflearningofficer.com/2025/01/17/the-reskilling-roadmap-navigati...↩
8. How to Keep Up with AI Through Reskilling - Professional & Executive Development, accessed June 8, 2025, https://professional.dce.harvard.edu/blog/how-to-keep-up-with-ai-through-reski...↩
9. AI Education Partnerships | TEEI – The Educational Equality Institute, accessed June 8, 2025, https://theeducationalequalityinstitute.org/ai-education-partnerships/↩
10. Initiative on AI and Education - The Spencer Foundation, accessed June 8, 2025, https://www.spencer.org/initiative-on-ai-and-education↩
11. The role of policies on technology and AI for innovation and increased competitiveness in North America - Brookings Institution, accessed June 8, 2025, https://www.brookings.edu/articles/the-role-of-policies-on-technology-and-ai-fo...↩
12. Brookings AI Policy Idea Incubator | Brookings - Brookings Institution, accessed June 8, 2025, https://www.brookings.edu/collection/brookings-ai-policy-idea-incubator/↩
13. Artificial Intelligence | NSF - National Science Foundation, accessed June 8, 2025, https://www.nsf.gov/focus-areas/artificial-intelligence↩
14. Advancing Artificial Intelligence Education for American Youth - The White House, accessed June 8, 2025, https://www.whitehouse.gov/presidential-actions/2025/04/advancing-artificial-int...↩
15. How AI is Transforming Nonprofits in 2025 – Giving USA, accessed June 8, 2025, https://givingusa.org/how-ai-is-transforming-nonprofits-in-2025/↩
16. Home - AI for Good Foundation - AI for social impact, accessed June 8, 2025, https://ai4good.org/↩
17. Labor Organizing and AI: The Employer Perspective | Littler, accessed June 8, 2025, https://www.littler.com/news-analysis/asap/labor-organizing-and-ai-employer-pe...↩
18. AI and the Future for Workers: Labor Union Responses to Artificial Intelligence, accessed June 8, 2025, https://www.ilr.cornell.edu/carow/events/ai-and-future-workers-labor-union-resp...↩
19. Federal AI legislation: An evaluation of existing proposals and a road map forward, accessed June 8, 2025, https://www.epi.org/publication/federal-ai-legislation/↩
20. How Government Can Embrace AI and Workers | Urban Institute, accessed June 8, 2025, https://www.urban.org/urban-wire/how-government-can-embrace-ai-and-worke...↩
21. Gen-AI: Artificial Intelligence and the Future of Work - IMF, accessed June 8, 2025, https://www.imf.org/-/media/Files/Publications/SDN/2024/English/SDNEA2024001.ashx↩
22. Q&A: What is the productivity gap? | Blog | Economics - Tutor2u, accessed June 8, 2025, https://www.tutor2u.net/economics/blog/qa-what-is-the-productivity-gap↩
23. What is Productivity Gap Metrics - Prohance.net, accessed June 8, 2025, https://www.prohance.net/glossary/what-is-productivity-gap-metrics.php↩
24. Production Gap: What It Is, How It Works, FAQs - Investopedia, accessed June 8, 2025, https://www.investopedia.com/terms/p/production-gap.asp↩
25. www.epi.org, accessed June 8, 2025, https://www.epi.org/blog/growing-inequalities-reflecting-growing-employer-pow...↩
26. Productivity | Definition, Measurement, Types, Examples, & Facts | Britannica Money, accessed June 8, 2025, https://www.britannica.com/money/productivity/Historical-trends↩
27. Artificial Intelligence Impact on Labor Markets - International Economic Development Council (IEDC), accessed June 8, 2025, https://www.iedconline.org/clientuploads/EDRP%20Logos/AI_Impact_on_Labor_M...↩
28. AI could make us more productive, can it also make us better paid? | World Economic Forum, accessed June 8, 2025, https://www.weforum.org/stories/2025/05/productivity-pay-artificial-intelligence/↩
29. AI productivity gains may be smaller than you're expecting | articles ..., accessed June 8, 2025, https://think.ing.com/articles/macro-level-productivity-gains-ai-coming-artificial-...↩
30. Artificial Intelligence and the Modern Productivity Paradox: A Clash ..., accessed June 8, 2025, https://www.nber.org/system/files/working_papers/w24001/w24001.pdf↩
31. The impact of artificial intelligence on macroeconomic productivity – CEPR, accessed June 8, 2025, https://cepr.org/voxeu/columns/impact-artificial-intelligence-macroeconomic-pro...↩
32. Generative AI is not replacing jobs or hurting wages at all, economists claim - The Register, accessed June 8, 2025, https://www.theregister.com/2025/04/29/generative_ai_no_effect_jobs_wages/↩
33. Generative AI isn't biting into wages, replacing workers, and isn't saving time, economists say | TechRadar, accessed June 8, 2025, https://www.techradar.com/pro/generative-ai-isnt-biting-into-wages-replacing-w...↩
34. AI's journey through time and technology across generations - Solveo, accessed June 8, 2025, https://solveo.co/ais-journey-through-time-and-technology-across-generations/↩
35. Internet and AI: similarities 22 years apart! - ExoB2B, accessed June 8, 2025, https://exob2b.com/en/internet-and-ai-similarities-22-years-apart/↩
36. "AI will replace all the jobs!" Is Just Tech Execs Doing Marketing - SparkToro, accessed June 8, 2025, https://sparktoro.com/blog/ai-will-replace-all-the-jobs-is-just-tech-execs-doing-...↩
37. The Productivity Revolution - Hoover Institution, accessed June 8, 2025, https://www.hoover.org/research/productivity-revolution↩
38. Access to Computer Technologies at Home Improves Wages in the Marketplace - CSUSB ScholarWorks, accessed June 8, 2025, https://scholarworks.lib.csusb.edu/cgi/viewcontent.cgi?article=1174&context=jitim↩
39. Superagency in the workplace: Empowering people to unlock AI's full potential - McKinsey, accessed June 8, 2025, https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/superagenc...↩
40. Three Reasons Why AI May Widen Global Inequality | Center For Global Development, accessed June 8, 2025, https://www.cgdev.org/blog/three-reasons-why-ai-may-widen-global-inequality↩
41. Introducing the Anthropic Economic Index, accessed June 8, 2025, https://www.anthropic.com/news/the-anthropic-economic-index↩
42. Reskilling the Workforce With AI: Harvard Business School's Raffaella Sadun, accessed June 8, 2025, https://sloanreview.mit.edu/audio/reskilling-the-workforce-with-ai-harvard-busin...↩