The Unemployed Majority? AI Displacement, Precarity, and Political Volatility

Full report

On this page

Introduction: The Productivity Paradox and the Crisis of Security

In the contemporary macroeconomic landscape, a profound paradox has emerged: human society has never possessed greater technological capability or generated more aggregate wealth, yet the material and psychological security of the ordinary worker appears increasingly fragile. The widespread deployment of Generative Artificial Intelligence (GenAI), advanced algorithmic management systems, and machine learning models has catalyzed fears of a structural collapse in labor demand. Popular discourse frequently centers on the specter of absolute mass unemployment—a dystopian future where capital owners reap the limitless yields of artificial general intelligence (AGI) while the human workforce is rendered permanently obsolete. However, a rigorous examination of labor economics, occupational exposure data, and sociological trends suggests a different, perhaps more insidious, trajectory. The immediate threat of the AI revolution is not necessarily the overnight elimination of all human labor, but rather the systemic degradation of job quality, the compression of career trajectories, and the severe erosion of worker bargaining power. The International Labour Organization (ILO) currently estimates that roughly one in four workers globally occupy jobs with some GenAI exposure, yet emphasizes that partial transformation of tasks is far more likely than outright occupational replacement for the vast majority of the workforce1. The defining economic crisis of the coming decade will likely not be mass idleness, but pervasive precarity. When entry-level roles disappear, when human labor is subjected to the relentless pacing of algorithmic surveillance, and when the returns on higher education evaporate in a saturated market for credentialed elites, the social contract inevitably fractures. This report provides an exhaustive analysis of the intersection between AI-driven labor displacement, the rise of the educated precariat, and the subsequent political volatility. By integrating the task-based macroeconomic models of automation, empirical data on observed AI deployment, structural-demographic theories of elite overproduction, and the political economy of status anxiety, this analysis answers the era's most explosive question: If society has never been more productive, why are ordinary lives less secure?

1. The Macroeconomics of AI and Task Transformation

To understand the labor market impacts of GenAI, it is necessary to abandon the binary framework of "employed" versus "replaced." The modern economy does not consist of indivisible jobs, but rather bundles of discrete tasks. Evaluating AI through a task-based framework reveals profound disparities in macroeconomic forecasts, pitting exuberant Wall Street projections against more sober academic analyses.

1.1 Conflicting Forecasts: The GDP and Productivity Debate

The rapid emergence of Large Language Models (LLMs) has generated highly divergent macroeconomic forecasts. Institutions such as Goldman Sachs have projected that generative AI could deliver a 7% increase in global GDP over a ten-year period, representing a roughly \$7 trillion boost to the global economy driven by widespread labor productivity gains3. These models largely assume that the automation of cognitive tasks will seamlessly translate into higher output and the reallocation of human capital to higher-value activities. Conversely, rigorous macroeconomic modeling by economists such as Daron Acemoglu presents a markedly more constrained outlook. In the 2024 NBER working paper The Simple Macroeconomics of AI, Acemoglu calculates that the total factor productivity (TFP) gains from AI over the next decade will likely be "nontrivial, but modest," translating to a maximum GDP growth of approximately 1.8%, with a more realistic baseline of just 0.71% to 1.1%3. This discrepancy arises from the distinction between automation and total factor productivity. Acemoglu and Restrepo’s foundational task-based framework demonstrates that technology affects labor demand by altering the task content of production. When automation occurs, it allows capital to take over tasks previously performed by humans, generating a "displacement effect" that inherently reduces the labor share of value-added income8. Historically, the labor share remained stable because previous technological waves simultaneously generated a "reinstatement effect"—creating entirely new, labor-intensive tasks that absorbed displaced workers7. The current precarity stems from the deployment of what the literature terms "so-so technologies." These are automated systems that are just productive enough to displace human labor and justify adoption by cost-cutting firms, but fail to generate the massive productivity surges required to spur broader economic growth and new task creation7. Consequently, the displacement effect dominates, wages stagnate relative to output, and the labor share of national income declines, severing the traditional link between technological progress and broad-based prosperity8.

1.2 Global Exposure and Demographic Disparities

The May 2025 ILO-NASK Global Index of Occupational Exposure provides the most rigorous empirical baseline for this phenomenon, assessing nearly 30,000 real-world job descriptions. The report confirms that 25% of global employment falls into occupations potentially exposed to GenAI2. However, this exposure is radically stratified by geography, income level, and gender.

The gender disparity is particularly acute. The ILO notes that in high-income countries, the share of female employment at the highest risk of AI-driven task automation is nearly three times that of men2. This is driven by the structural overrepresentation of women in clerical, administrative, and data-processing roles. Generative AI is exceptionally adept at the routine cognitive tasks that define these occupations—document formatting, scheduling, and data entry. While these roles will not vanish entirely, partial automation threatens to hollow them out, reducing required skill levels, stagnating wages, and increasing vulnerability to algorithmic micromanagement without offering targeted pathways for role redesign2.

2. Theoretical vs. Observed Exposure and the Deployment Gap

While theoretical exposure to AI is vast, real-world macroeconomic disruption is currently gated by significant deployment frictions. A landmark 2026 study by Anthropic researchers Massenkoff and McCrory introduced the metric of "Observed Exposure," measuring actual enterprise usage of large language models against theoretical capabilities16.

2.1 The Deployment Gap

The researchers discovered a massive "deployment gap." By comparing theoretical task vulnerability to billions of anonymized enterprise AI queries, the study found that AI is being utilized far more narrowly than capability models suggest16. Institutional inertia, legal constraints, data privacy requirements, and the necessity for human verification have dramatically slowed the translation of theoretical capability into deployed automation16. For example, in Computer and Mathematical occupations, theoretical models suggest that 94% of tasks are feasible for LLM automation. However, actual observed coverage in 2026 is only 33%, representing a 61 percentage point gap. Similarly, Office and Administrative roles exhibit a 90% theoretical exposure but roughly 25% observed exposure18. At the opposite end of the spectrum, approximately 30% of the US workforce—including cooks, mechanics, lifeguards, and general repair workers—shows absolutely zero observed AI exposure in their daily tasks, as physical and contextual judgment remains entirely outside current LLM capabilities16.

2.2 The Demographics of the Exposed

The Anthropic data radically challenges early assumptions about who AI threatens. Historically, automation targeted routine manual labor. Generative AI targets the apex of the knowledge economy. Workers in the most highly exposed AI occupations exhibit a distinct demographic profile:

• They earn, on average, 47% higher hourly wages (\$32.69 vs. \$22.23 for the zero-exposure group)16.
• They are 16 percentage points more likely to be female (54.4% vs. 38.8%)20.
• They are nearly four times more likely to hold a graduate degree (17.4% vs. 4.5%)16.
• They are disproportionately represented in roles such as Computer Programmers (74.5% observed exposure), Customer Service Representatives (70.1%), and Financial Analysts (57.2%)20.

2.3 The Silent Disruption: The Youth Hiring Chill

Perhaps the most critical finding from the Anthropic deployment data is the absence of a mass layoff event, coupled with a highly concerning leading indicator. The researchers found no systemic spike in unemployment for highly exposed workers since the launch of ChatGPT in late 2022; the economy has not buckled under mass terminations16. Instead, the disruption is manifesting as a structural "hiring chill." The data reveals a 14% drop in the hiring rate of young workers (aged 22–25) into AI-exposed occupations18. Firms are not firing their senior engineers, lawyers, or analysts; rather, they are utilizing AI tools to dramatically increase the productivity of their existing senior staff, thereby eliminating the need to hire junior talent to handle routine foundational tasks13. This phenomenon signals a profound shift in workforce development: the eradication of the entry-level career stage.

3. The Attrition of the Entry Level and the "AI Chain"

The 14% contraction in youth hiring across exposed sectors points to a destabilizing long-term consequence of AI integration. Historically, entry-level roles—whether junior paralegals drafting briefs, associate developers writing boilerplate code, or financial analysts building base models—served dual economic functions. For the firm, they provided cost-effective labor for data processing and verification. For the worker, they served as paid apprenticeships, providing the informal environments required to acquire tacit industry knowledge, operational judgment, and social capital25.

3.1 The Theory of AI Automation and Task Chaining

The mechanics of this entry-level displacement are formally modeled in the 2026 NBER working paper Chaining Tasks, Redefining Work: A Theory of AI Automation by Demirer, Horton, Immorlica, Lucier, and Shahidi. The authors note that production is inherently a sequence of steps. Historically, dividing a job among multiple human workers incurred high coordination costs (handoffs, communication, managing dependencies), which dictated how firms bounded job roles and limited excessive specialization26. AI fundamentally alters this calculus. When an AI agent executes contiguous steps in a workflow (forming an "AI chain"), the coordination costs between those specific steps drop effectively to zero. Crucially, a human worker is only required at the end of the chain to verify the final output26. Because human verification is reduced to a fixed cost at the terminus of a process rather than a marginal cost scaling with every step, firms are heavily incentivized to cluster AI-exposed steps and absorb as many contiguous junior-level tasks as possible into automated chains26. The empirical evidence supports this: AI-executed steps co-occur in chains, and adjacency to an AI-executed step drastically increases the likelihood that a neighboring task will also be automated26.

3.2 Knowledge Collapse and the Erosion of Human Capital

The result of task chaining is a structural hollowing out of the firm. The tasks that once comprised the daily workflow of a junior employee are now subsumed into the AI chain operated by a senior worker. The senior worker becomes vastly more productive, but the pathway to becoming a senior worker is destroyed19. When entry-level roles disappear, a macroeconomic vulnerability emerges: the economy faces a paradox of simultaneous labor shortages in senior roles and a surplus of blocked, underutilized youth25. If AI absorbs the foundational, repetitive tasks through which humans build neural pathways and domain expertise, the workforce faces a looming crisis of competence. This systemic risk is formalized in the 2026 NBER paper AI, Human Cognition and Knowledge Collapse by Acemoglu, Kong, and Ozdaglar. The authors model how the deployment of generative AI—and specifically autonomous "agentic AI"—shapes human learning incentives29. In their dynamic DeGroot learning model, societal general knowledge relies on the aggregate learning efforts of individuals exploring idiosyncratic contexts. However, highly accurate agentic AI acts as a substitute for human cognitive effort. Because individual agents rationally choose to rely on the AI rather than exert effort to learn, their private learning diminishes30. The paradox is that community-level knowledge is the very training data that future AI models rely upon. The authors demonstrate that if human effort drops below a critical threshold, the flow of new information ceases, triggering a "knowledge collapse"30. If a community's stock of general knowledge dips below this tipping point, learning incentives and collective knowledge spiral downward together, rendering both human workers and the AI systems that rely on them fundamentally less capable over the long run31.

4. Algorithmic Management, Deskilling, and Bargaining Power

Where human labor remains necessary, its nature is being fundamentally altered. The deployment of AI is not solely directed at cognitive task automation; it is increasingly directed at labor control. "Algorithmic management" (AM) refers to the use of AI to automate traditional managerial functions: monitoring, scheduling, pacing, evaluating, and disciplining workers33.

4.1 The Expansion of the Algorithmic Panopticon

Originally confined to the platform and gig economy (e.g., ride-sharing, food delivery), algorithmic management has thoroughly penetrated logistics, warehousing, retail, and traditional white-collar office environments35. AI systems utilize keystroke loggers, biometric sensors, hard-hat cameras, and real-time data analytics to track physical movements and cognitive outputs continuously38. The primary economic function of algorithmic management is the extraction of relative surplus value by minimizing "idle time" and accelerating the pace of work34. In logistics and retail, algorithmic shift scheduling and automated work method instructions strip workers of their autonomy. This process drives pervasive deskilling. By standardizing tasks into micro-instructions and removing the need for human judgment, employers effectively commodify the labor pool. This makes workers easily replaceable, weakening internal career ladders and driving down the wage premium associated with experience36.

4.2 The Shift in Structural Power

Algorithmic management represents a massive transfer of structural power from labor to capital. The opacity of automated decision-making systems creates severe information asymmetries36. Workers are subjected to "unstable algorithmic accountability," where the rules of success fluctuate invisibly based on shifting optimization goals, generating intense psychological stress, status loss, and occupational burnout33. The fundamental reality of AI deployment is shaped by power dynamics: the less structural power a class of workers possesses, the more aggressively AI is weaponized to erode their job quality and wages38.

4.3 Regulatory Responses and the Resurgence of Collective Bargaining

The scale of this power imbalance has triggered institutional responses, particularly in Europe. The European Union's Platform Work Directive (PWD) and the AI Act attempt to introduce guardrails, emphasizing the "human in command" principle, mandating transparency in automated decision-making, and requiring fundamental rights impact assessments33. However, legal scholars and labor advocates note that the AI Act's market-oriented design largely overlooks the specific power imbalances inherent to the employer-employee relationship, leaving vast loopholes for private employers outside the gig economy33. Consequently, the most effective resistance has emerged through collective bargaining. Organizations like UNI Europa have established databases tracking AI clauses negotiated by trade unions and works councils44. These agreements focus on co-determination, mandating that employers consult worker representatives before deploying AI, guaranteeing the right to human review of algorithmic decisions, restricting the use of workforce analytics for disciplinary purposes, and establishing "digi-road maps" to ensure technology augments rather than surveils the workforce37. In the United States, the historic 2023 strikes by the Writers Guild of America (WGA) and the Screen Actors Guild (SAG-AFTRA) represented a watershed moment. The unions recognized that generative AI threatened to convert their specialized expertise into a low-cost commodity. By leveraging their collective power, they successfully negotiated contracts that strictly govern the use of generative models, ensuring that technology remains a tool utilized by human creators rather than an autonomous replacement40. These victories highlight that the trajectory of AI adoption is not a technological inevitability, but a contested political choice.

5. Regional Vulnerability: The Cicero, Illinois Case Study

The macroeconomic pressures of AI displacement, task chaining, and algorithmic management do not occur in an abstract vacuum; they manifest sharply in specific regional geographies. The town of Cicero, Illinois, serves as a vital case study for understanding how technological disruption intersects with demographic vulnerability and the hollowing out of the industrial middle class. Cicero, bordering Chicago, is a densely populated municipality of approximately 85,000 residents, characterized by a predominantly Hispanic and Latino population (87.5%) and a heavy reliance on blue-collar and mid-tier service employment49.

While current LLM technology primarily threatens knowledge work, the broader ecosystem of automation—including AI-driven robotics, computer vision, and algorithmic logistics—poses an existential threat to regional hubs like Cicero. Illinois maintains a manufacturing workforce that is 16% more concentrated than the national average, heavily driven by chemical manufacturing, transportation, and material moving51. However, the regional labor market is deeply stratified. In the broader Chicago area, out-of-school and out-of-work rates for young Black men (36.6%) and young Latino men (19.8%) dwarf those of young white men (5.3%)53. While Latino workers have maintained strong access to manufacturing roles through community networks, the sector is aging rapidly; manufacturing has the smallest share of workers aged 14 to 34 and the largest share of workers over 4553. As AI and advanced robotics scale, the physical and logistical tasks that form the backbone of Cicero’s economy are highly exposed to algorithmic management and eventual automation. The deployment of automated scheduling and warehouse pace-setting disproportionately impacts immigrant and minority workforces, exacerbating regional inequality38. For communities like Cicero, where only 10.8% of the population holds a bachelor's degree, the destruction of the entry-level manufacturing and logistics pipeline removes one of the few remaining accessible pathways to the middle class, cementing a regional "middle-income trap"13.

6. Elite Overproduction and the Educated Precariat

The disruption of the working-class industrial base is now running parallel to a crisis at the upper echelons of the economy. The macroeconomic pressures of AI displacement are interacting with long-term structural-demographic trends to produce a highly combustible social environment, driven by the emergence of the "educated precariat."

6.1 The Cliodynamics of Elite Overproduction

Peter Turchin’s theory of cliodynamics provides a vital framework for understanding contemporary political instability. Turchin identifies "elite overproduction" as the primary driver of societal breakdown. Over recent decades, advanced economies have functioned as "wealth pumps," funneling economic gains upward while simultaneously expanding higher education to produce vast cohorts of credentialed graduates55. The result is a structural imbalance: society produces far more "elite aspirants" (individuals with advanced degrees expecting high-status, high-income careers) than there are actual elite positions available in the economy or the political sphere55. Turchin uses the metaphor of the 535 seats in the US Congress; as the number of wealthy and credentialed aspirants balloons, the supply of power remains fixed, forcing a brutal, zero-sum intra-elite competition that splinters society into warring factions55.

6.2 AI as a Catalyst for Precarity

The advent of Generative AI acts as a supercharger to this elite overproduction dynamic. As noted in the Anthropic data, AI currently targets the apex of the knowledge economy—the very sectors (software, finance, law, media) that elite aspirants rely upon16. As AI task-chaining eliminates the entry-level roles that traditionally served as the on-ramps to elite status, the pipeline of upward mobility is severed23. The generation entering the workforce today finds themselves highly educated, heavily indebted, and structurally locked out of the economic security enjoyed by previous generations. This precarity is compounded by severe housing cost burdens, which prevent young professionals from achieving traditional milestones of adulthood, such as homeownership and family formation58. When a society produces an overabundance of credentialed youths but denies them the economic agency and status they were promised, the historical outcome is almost universally political rupture. Turchin's models suggest that when societies reach this level of structural stress, they face an 85% probability of severe political violence or state collapse unless a massive "integrative" policy pivot turns off the wealth pump56.

7. Status Anxiety and the Politics of Volatility

This economic precarity translates directly into psychological distress and political radicalization through the mechanism of "status anxiety." As detailed by sociologists Noam Gidron and Peter A. Hall, the subjective experience of downward social mobility—or the pervasive fear of it—is a potent catalyst for political volatility60. Automation and algorithmic deskilling do not merely threaten income; they threaten occupational identity and societal dignity63. When workers perceive that their hard-earned skills are being devalued and their societal standing is slipping, they experience acute status anxiety. Frustrated populations, feeling abandoned by mainstream institutions that champion unchecked technological progress, increasingly turn to anti-establishment populism. Populist rhetoric effectively weaponizes this resentment, framing the economic marginalization of the working and middle classes as a deliberate betrayal by a disconnected cultural and economic elite60. Empirical research indicates a direct correlation between occupational exposure to automation and increased support for both left-wing and right-wing populist movements65. Right-wing populism, in particular, proves highly effective at channeling status anxiety by promising to restore traditional hierarchies and an idealized past, linking economic grievances with cultural and nationalistic backlash61.

7.1 Automation, Inequality, and the Threat of Repression

The concentration of wealth and power generated by unmitigated AI automation poses an existential threat to democratic institutions. In their 2026 NBER working paper Automation and Repression, Acemoglu, Gitmez, and Shadmehr explore the political game theory of extreme inequality. In their model, as capital accumulates and automation displaces human labor, the economic chasm between capital owners and the working class widens, predictably increasing the likelihood of civil unrest and worker revolts67. Historically, elites faced with the threat of mass revolt had two options: redistribute wealth (often through the expansion of the welfare state) or invest in state repression68. The chilling conclusion of their mathematical model is that there is a deep complementarity between automation and repression. As the capital share grows massive and labor becomes economically obsolete, elites find it economically preferable to fund authoritarian security apparatuses and suppress democratic feedback mechanisms rather than surrender their vast tech-generated wealth to taxation and redistribution68. If the structural wealth pump remains active, and millions of downwardly mobile, educated youths are left to face a precarious gig-economy reality, the probability of severe domestic conflict skyrockets56. History demonstrates that legitimacy is perishable; when a critical mass of the population determines that the economic rules are fundamentally rigged, the social order violently resets56.

8. Forging a Pro-Worker AI and Post-Work Safety Net

Averting political disintegration requires radical, proactive policy interventions. The solution cannot merely be reactive compensation; it must involve reshaping the trajectory of technological development and reimagining the macroeconomic safety net.

8.1 Steering Technology: Building Pro-Worker AI

The most critical intervention is altering the incentive structures that dictate how AI is designed and deployed. In Building Pro-Worker Artificial Intelligence (NBER 2026), Acemoglu, Autor, and Johnson argue that the current focus on mimicking and automating human expertise is an ideological and institutional choice, not a technological inevitability39. The authors advocate for a shift toward "expertise-leveling" and "labor-augmenting" AI39. Rather than using AI to replace the software developer or the junior analyst, technology should be designed to expand the capabilities of frontline workers—providing nurses with advanced diagnostic tools, or empowering tradespeople to perform complex engineering assessments39. By democratizing specialized knowledge, AI can increase the marginal productivity and market value of human labor, fostering a new middle class rather than annihilating the existing one39. Achieving this requires specific policy levers:

• Tax Equalization: Eliminating the current tax bias that heavily favors capital investments (software and automation) over human labor (payroll taxes), effectively ending the artificial subsidy for worker replacement74.
• Antitrust Enforcement: Breaking the monopolistic control of Big Tech over AI research, ensuring a diversity of business models that explore human complementarity rather than pure automation39.
• Institutionalizing Worker Voice: Mandating co-determination in technology adoption, ensuring that those subjected to AI tools have a structural role in their design and implementation, utilizing the WGA and SAG-AFTRA strikes as a legislative blueprint40.

8.2 UBI vs. UBS: Redesigning the Safety Net

Even with pro-worker AI, the structural reduction in aggregate labor hours necessitates a fundamental redesign of income distribution. As the link between traditional employment and survival weakens, two competing macroeconomic stabilization models have gained prominence: Universal Basic Income (UBI) and Universal Basic Services (UBS)13.

While often debated as mutually exclusive, leading economists increasingly view UBI and UBS as complementary pillars required to transition to a highly automated economy80. The elasticity of substitution between capital and labor dictates that as AI scales globally at a negligible marginal cost, the labor share of income will continue to fall13. By implementing a robust UBI funded by progressive taxes on capital and corporate profits, alongside expansive UBS, the state can effectively sever the existential terror of unemployment from the reality of technological displacement. This dual approach neutralizes the macroeconomic demand constraint, providing the purchasing power necessary to keep the consumption-driven economy functioning while simultaneously restoring human dignity and autonomy13.

Conclusion

The premise of the "Unemployed Majority" captures the anxiety of the moment, but misdiagnoses the immediate pathology. The primary threat of the AI era is not a sudden, absolute absence of work. It is the systemic hollowing out of the labor market—the chaining of entry-level tasks that destroys career mobility, the imposition of algorithmic management that strips labor of autonomy and dignity, and the deployment of "so-so" automation that enriches capital while stagnating wages. We are currently navigating a perilous divergence. Society possesses the technological capacity to generate unprecedented abundance, yet the institutional mechanisms required to distribute that abundance are paralyzed by an ideology that favors automation over augmentation. The result is an economy defined by elite overproduction, where highly educated youths face blocked trajectories and acute status anxiety, while regional manufacturing hubs face the silent attrition of the middle-income trap. As these economic realities translate into political grievances, the foundations of democratic stability are shaken, opening the door to populism, polarization, and ultimately, elite repression. Technology is not an autonomous force of nature; it is a manifestation of social and political choices. The precarity of the modern worker is the intended output of an economic architecture that prioritizes the automation of expertise and the surveillance of labor. To ensure that the dividends of Artificial Intelligence facilitate societal flourishing rather than political disintegration, structural power must be counterbalanced. Through the reinvigoration of organized labor, the aggressive legislative steering of AI development toward human augmentation, and the establishment of unconditional economic floors like Universal Basic Income and Services, society can resolve the productivity paradox and reclaim the future of work.

Works cited

1. Generative AI and jobs: a refined global index of occupational, https\://www\.researchgate.net/publication/391951420\_Generative\_AI\_and\_jobs\_a\_refined\_global\_index\_of\_occupational\_exposure

2. AI threatens one in four jobs – but transformation, not replacement, https\://news.un.org/en/story/2025/05/1163486

3. A new look at the economics of AI | MIT Sloan, https\://mitsloan.mit.edu/ideas-made-to-matter/a-new-look-economics-ai

4. How Will AI Impact the Labor Market? \- Goldman Sachs, https\://www\.goldmansachs.com/insights/goldman-sachs-exchanges/how-will-ai-impact-the-labor-market

5. The Potentially Large Effects of Artificial Intelligence on Economic, https\://www\.gspublishing.com/content/research/en/reports/2023/03/27/d64e052b-0f6e-45d7-967b-d7be35fabd16.html

6. The Simple Macroeconomics of AI | MIT Economics, https\://economics.mit.edu/sites/default/files/2024-04/The%20Simple%20Macroeconomics%20of%20AI.pdf

7. The simple macroeconomics of AI\* \- MIT Economics, https\://economics.mit.edu/sites/default/files/2024-10/The%20Simple%20Macroeconomics%20of%20AI.pdf

8. acemoglu-restrepo-2019-automation-and-new-tasks-how, https\://shapingwork.mit.edu/wp-content/uploads/2023/10/acemoglu-restrepo-2019-automation-and-new-tasks-how-technology-displaces-and-reinstates-labor.pdf

9. Automation's Impact on Labor Demand | PDF | Labour Economics, https\://www\.scribd.com/document/864637972/Acemoglu-Restrepo-2019-JEP

10. Acemoglu & Restrepo, The Task Model of the Labor Share, https\://www\.mikekonczal.com/blogs\_2026\_07\_AI\_uses/reading\_companion\_AR\_labor\_share.html

11. Tasks At Work: Comparative Advantage, Technology and Labor, https\://www\.nber.org/papers/w32872

12. The Work of the Future: \- Shaping Technology and Institutions \- MIT, https\://web.mit.edu/2.810/www/files/readings/WorkOfTheFuture.pdf

13. AI Job Loss: Will A Universal Basic Income Be Required?, https\://www\.theemploymentlawsolicitors.co.uk/news/2026/02/24/ai-job-loss/

14. One in four jobs at risk of being transformed by GenAI, new ILO, https\://www\.ilo.org/resource/news/one-four-jobs-risk-being-transformed-genai-new-ilo%E2%80%93nask-global-index-shows

15. 25% of global employment at risk of exposure to GenAI, https\://guardian.ng/appointments/25-of-global-employment-at-risk-of-exposure-to-genai/

16. Labor market impacts of AI: A new measure \- Anthropic, https\://www\.anthropic.com/research/labor-market-impacts

17. Labor Market Impacts of AI: A New Measure and Early Evidence, https\://policycommons.net/artifacts/46350186/a42bc3fc08283562f08fd8bdee8f6f9a3d506e87/47248964/

18. Which Sales Jobs AI Replaces: Anthropic 2026 Data, https\://r-sun.ai/insights/ai-labor-market-sales-tech

19. AI and the Labor Market: What the Data Actually Shows | Alluxi, https\://www\.alluxi.com/blog/detail/ai-and-the-labor-market-what-the-data-actually-shows

20. AI Job Displacement Data Exposes a \$1 Trillion Due Diligence Blind, http\://developmentcorporate.com/saas/ai-job-displacement-data-exposes-a-1-trillion-due-diligence-blind-spot/

21. Maxim Massenkoff | Homepage, https\://maximmassenkoff.com/

22. AI job exposure by occupation: 800 roles ranked for 2026 | PayScope, https\://www\.payscope.ai/blog/ai-job-exposure-by-occupation-2026

23. AI Is Replacing 42 Million Workers. Are You One of Them? \- Medium, https\://medium.com/@srivatsavgopinath2000/ai-is-replacing-42-million-workers-are-you-one-of-them-58fd3191d649

24. AI's real labor impact: The hiring chill hits young people \- IO+, https\://ioplus.nl/en/posts/ais-real-labor-impact-the-hiring-chill-hits-young-people

25. When Platforms Replace the Pipeline: AI, Labor Erosion, and ... \- MDPI, https\://www\.mdpi.com/2813-4176/4/3/11

26. Chaining Tasks, Redefining Work: A Theory of AI Automation \- NBER, https\://www\.nber.org/system/files/working\_papers/w34859/w34859.pdf

27. AI in Finance and Banking, March 15, 2026 \- LLRX, https\://www\.llrx.com/2026/03/ai-in-finance-and-banking-march-15-2026/

28. Teaming Up with AI: Coordination and Cooperation \- arXiv, https\://arxiv.org/pdf/2607.03181

29. AI, Human Cognition and Knowledge Collapse | NBER, https\://www\.nber.org/papers/w34910

30. AI, Human Cognition and Knowledge Collapse | MIT Economics, https\://economics.mit.edu/sites/default/files/2026-02/AI%2C%20Human%20Cognition%20and%20Knowledge%20Collapse%2002-20-26.pdf

31. AI, Human Cognition and Knowledge Collapse | MIT Economics, https\://economics.mit.edu/sites/default/files/2026-05/AI%2C%20Human%20Cognition%20and%20Knowledge%20Collapse%2005-05-26.pdf

32. Research Insight: Will AI lead to the collapse of collective knowledge?, https\://shapingwork.mit.edu/research/research-insight-will-ai-lead-to-the-collapse-of-collective-knowledge/

33. Digitalisation, artificial intelligence and algorithmic management in, https\://www\.europarl.europa.eu/RegData/etudes/STUD/2025/774670/EPRS\_STU(2025)774670\_EN.pdf

34. The impact of artificial intelligence on the nature and quality of jobs, https\://www\.econstor.eu/bitstream/10419/270468/1/1817796259.pdf

35. Brussels, 20.7.2026 SWD(2026) 239 final COMMISSION STAFF, https\://employment-social-affairs.ec.europa.eu/document/download/a3bad071-c210-4dba-b632-c8737e7f7eff\_en?filename=SWD\_2026\_239\_1\_EN\_autre\_document\_travail\_service\_part1\_v3.pdf

36. Worker voice and algorithmic management in post-Brexit Britain, https\://www\.researchgate.net/publication/365780882\_Worker\_voice\_and\_algorithmic\_management\_in\_post-Brexit\_Britain

37. Algorithmic Management: Opportunities for Collective Action. A, https\://uniglobalunion.org/wp-content/uploads/Algorithmic-Management-Opportunities-for-Collective-Action.pdf

38. AI surveillance is further exploiting U.S. labor. Here's how to reclaim, https\://racepowerpolicy.org/2026/03/ai-surveillance-and-exploitation-are-pushing-down-labor-value/

39. Building Pro-Worker Artificial Intelligence, https\://www\.nber.org/system/files/working\_papers/w34854/w34854.pdf

40. Boosting U.S. worker power and voice in the AI-enabled workplace, https\://equitablegrowth.org/boosting-u-s-worker-power-and-voice-in-the-ai-enabled-workplace/

41. WORKING PAPER \- EUI Cadmus, https\://cadmus.eui.eu/server/api/core/bitstreams/afce372a-6a5f-4299-aca0-38e8dde2ba88/content

42. Unstable algorithmic accountability and non-algorithmic labour, https\://ephemerajournal.org/contribution/unstable-algorithmic-accountability-and-non-algorithmic-labour-problems-lithuanias-0

43. Algorithmic management and collective bargaining, https\://lirias.kuleuven.be/retrieve/722036

44. A database of AI and algorithmic management in collective, https\://www\.uni-europa.org/news/a-database-of-ai-and-algorithmic-management-in-collective-bargaining-agreements/

45. Collective Bargaining Practices on AI and Algorithmic Management, https\://www\.uni-europa.org/wp-content/uploads/sites/3/2024/03/CollectiveBargainingPracticesOnAI\_final.pdf

46. AI and algorithmic management in European services sectors., https\://collections.fes.de/publikationen/content/titleinfo/450342/full.pdf

47. Hollywood writers went on strike to protect their livelihoods from, https\://www\.brookings.edu/articles/hollywood-writers-went-on-strike-to-protect-their-livelihoods-from-generative-ai-their-remarkable-victory-matters-for-all-workers/

48. Governance of Generative AI in Creative Work: Consent, Credit, https\://ora.ox.ac.uk/objects/uuid:81fc314c-4c59-4a66-aac7-ea547fff1b21/files/r7m01bn28z

49. Cicero town, Illinois \- U.S. Census Bureau QuickFacts, https\://www\.census.gov/quickfacts/fact/table/cicerotownillinois/NES010223

50. Cicero, IL | Data USA, https\://datausa.io/profile/geo/cicero-il/

51. Chart of the Month: Illinois as a Manufacturing Destination—Key, https\://www\.illinoisedc.org/news/chart-of-the-month-illinois-as-a-manufacturing-destination-key-subsectors/

52. Chart of the Month: Illinois' Regional Workforce Dominance, https\://www\.illinoisedc.org/news/chart-of-the-month-illinois-regional-workforce-dominance/

53. Revitalizing Manufacturing and Expanding Opportunities for, https\://tcf.org/content/report/revitalizing-manufacturing-expanding-opportunities-chicagos-black-latino-communities/

54. Population loss, aging workforce threaten Illinois economy, https\://www\.illinoispolicy.org/population-loss-aging-workforce-threaten-illinois-economy/

55. End Times 2022044451, 2022044452, 9780593490501, https\://dokumen.pub/end-times-2022044451-2022044452-9780593490501-9780593490518.html

56. Elites Do Not Act with Enlightened Self-Interest Until it is Too Late, https\://medium.com/epoch-eclipse/elites-do-not-act-with-enlightened-self-interest-until-it-is-too-late-part-2-399356fb176e

57. What's the Solution to "Elite Overproduction" In Singapore \- Reddit, https\://www\.reddit.com/r/askSingapore/comments/1jg9cyp/whats\_the\_solution\_to\_elite\_overproduction\_in/

58. Reflections on Homeland Insecurity: The Strategic Anatomy of Civil, https\://www\.militarystrategymagazine.com/exclusives/reflections-on-homeland-insecurity-the-strategic-anatomy-of-civil-wars-to-come/

59. Conference Programme & Abstract Book: ACAH/ACCS/ACSS 2026, https\://issuu.com/iafor/docs/acah\_accs\_acss\_2026\_conference\_programme\_abstrac

60. University of Groningen Political Economy of Populism in Europe, https\://research.rug.nl/files/1295806823/Complete\_thesis.pdf

61. Explaining the Populist Right in the Neoliberal West \- MDPI, https\://www\.mdpi.com/2075-4698/13/5/110

62. Politicizing Status Loss Among Trump Supporters in 2020 | RSF, https\://www\.rsfjournal.org/content/8/6/69

63. Uncorrected Proof \- MIT Press Direct, https\://direct.mit.edu/euso/article-pdf/doi/10.1162/EUSO.a.100/2581621/euso.a.100.pdf

64. Competing demand-side explanations and populism: cross-national, https\://www\.tandfonline.com/doi/abs/10.1080/13510347.2024.2418457

65. (PDF) Did State Responses to Automation Matter for Voters?, https\://www\.researchgate.net/publication/332064850\_Did\_State\_Responses\_to\_Automation\_Matter\_for\_Voters

66. Sofia Mascarenhas Chumbo Pires Inflation, Living Costs, and, https\://repositorium.uminho.pt/bitstreams/bf50bcbd-d85a-4909-bec8-a0835d30780b/download

67. Automation and Repression \- NBER, https\://www\.nber.org/papers/w35336

68. Automation and Repression Daron Acemoglu, A. Arda Gitmez, and, https\://www\.nber.org/system/files/working\_papers/w35336/w35336.pdf

69. Mehdi Shadmehr | NBER, https\://www\.nber.org/people/mehdi\_shadmehr

70. AI News Briefing, June 21, 2026: D.A.D. Week In Review June 21, https\://dailyaidigest.net/archive/2026/06/21/

71. Building Pro-Worker Artificial Intelligence | NBER, https\://www\.nber.org/papers/w34854

72. Capabilities: The Power of AI to Alleviate Skill Shortages, https\://openknowledge.worldbank.org/bitstreams/cbbe027c-80f1-423d-b1fe-5f4e1a91824c/download

73. Daron Acemoglu \- Simon Johnson \- Power and Progress \- Scribd, https\://www\.scribd.com/document/716293014/Daron-Acemoglu-Simon-Johnson-Power-and-Progress-Our-Thousand-Year-Struggle-Over-Technology-and-Prosperity-Hachette-Book-Group-2023

74. AI and the Future of Labour Taxation \- Georg Kofler, https\://koflerge.org/wp-content/uploads/2026/04/kofler\_ai-and-labour-taxation.pdf

75. Pascual Restrepo \- NBER, https\://www\.nber.org/people/pascual\_restrepo

76. The Pope Should Have Gone Further on AI \- Social Europe, https\://www\.socialeurope.eu/the-pope-should-have-gone-further-on-ai

77. (PDF) The Need for Multipolar Artificial Intelligence Governance, https\://www\.researchgate.net/publication/393673006\_The\_Need\_for\_Multipolar\_Artificial\_Intelligence\_Governance

78. Universal basic income | Social Stratification Class Notes \- Fiveable, https\://fiveable.me/social-stratification/unit-11/universal-basic-income/study-guide/eLWuGC9TGRRa39cm

79. Basic Income as Common Dividends: Piloting a Transformative, https\://www\.progressiveeconomyforum.com/wp-content/uploads/2019/05/PEF\_Piloting\_Basic\_Income\_Guy\_Standing.pdf

80. Full article: Money for everything? Universal basic income in a crisis, https\://www\.tandfonline.com/doi/full/10.1080/03085147.2022.2035930

81. Post-Labor Economics: A Systematic Review \- Preprints.org, https\://www\.preprints.org/manuscript/202504.0444/v2

82. Economic Growth under Transformative AI Philip Trammell ... \- NBER, https\://www\.nber.org/system/files/working\_papers/w31815/revisions/w31815.rev1.pdf