Why the AI Readiness Narrative Misunderstands the Future of Work

AI readiness is a myth; AI is creating structural labor redundancy
A few AI-powered workers will replace the work of many
Policy must shift from training to managing disruption and inequality

The rapid advancements in large language models and automation are enabling them to perform tasks such as writing, designing, coding, advising, and managing at speeds previously unimaginable. This raises a key question: not just whether workers can adapt to AI, but whether there will be enough jobs requiring human labor in the future. Recent industry data indicate a rapid increase in the use of generative AI, with many firms now using it regularly (AI use by individuals surges across the OECD as adoption by firms continues to expand, 2026). Independent studies suggest that a significant portion of tasks in high-risk occupations, perhaps a quarter to a third, are at risk of automation (Xu et al., 2025). These two points highlight a difficult truth: the adoption of AI is a fundamental change that dramatically increases the output of a small group of workers, effectively replacing the work of many. This situation can be described as structural labor redundancy. Emphasizing policy readiness overlooks the large-scale, fast-moving nature of displacement, diverting resources from necessary changes in distribution, market structure, and institutions that are essential to determining who benefits and who is left behind.

The Issue of Structural Labor Redundancy and the Myth of 'Readiness'

Policymakers and well-intentioned experts often suggest investing in skills, improving digital knowledge, and increasing access to make the workforce AI-ready. While these investments are important, they do not fully address the mechanics of the new economy. AI does not just improve individual worker productivity by a small amount; it creates large advances, such as agents and models that can handle entire sets of tasks that previously required many people. When these technologies are integrated into companies, a single, highly competent individual can manage operations that would normally require entire teams (Giering & Kirchner, 2025). This does not lead to minor disturbances but to a major reshaping of demand. Employers will likely reduce their workforce to fewer AI-skilled positions, leading to a decline in mid-skill occupations, not because workers lack skills, but because companies can achieve much higher productivity with AI (Liu et al., 2024). This situation is what is referred to as structural labor redundancy.

This perspective is important because it changes how we approach policy. When readiness is set as a goal, it assumes that training many workers will allow them to fill improved roles. This structural redundancy indicates that the labor market will not be able to absorb many retrained workers, because there will be less need for human labor in those tasks. If a company has to choose between paying one person $150,000 to manage an AI setup that replaces 200 workers or paying 200 workers $40,000 each, it will often prefer the first option because of the scale and capital required (The Impact of Artificial Intelligence Adoption on Labor Cost Stickiness: Firm-Level Evidence from China, 2026). This is already happening as companies use generative AI to automate tasks such as customer support, content creation, and basic legal checks (Brynjolfsson et al., 2023, pp. 889-930). The policy error is thinking that readiness is enough. While it is needed, it is not enough. The major problem is how to distribute income, talent, and opportunities when AI concentrates production and rewards in a small group of the workforce.

Productivity Multipliers, Geographic Concentration, and Winner-Takes-All Dynamics

AI changes the way productivity works. Traditional automation replaced separate, manual tasks. Systems based on generative AI replace entire sets of tasks in thinking and administrative areas. Studies suggest that with rapid adoption, a large share of work hours could be automated by 2030 (Gordon & Murray, 2023). Companies that adopt these systems can greatly increase production while decreasing the number of employees. These gains are not evenly distributed; instead, they benefit companies that control models, data, computing resources, and skilled workers. This results in markets where a few top companies take the majority of the market share, profits, and talent.

This is happening not only in the corporate world, but also across different geographic regions. Strong AI activity tends to cluster in areas with a high concentration of talent, financial resources, and suitable infrastructure. China's recent industrial rollouts and the many models displayed at major conferences demonstrate how both state and private actors can build capacity in major centers. Information from recent global AI events showed that China alone has launched well over a thousand models, underscoring the advantages of dense clusters (News, 2024). The U.S. has several centers of innovation, but it is questionable whether it has enough backup locations to address challenges and distribute opportunities across different areas. Concentrated productivity leads to concentrated wages, meaning areas that cannot attract major AI companies will experience slow wage growth, talent loss, and little investment in local innovation.

These geographical factors also affect the labor market. If a small number of companies in a few cities can produce much more with fewer people, talent and money will gather around these hubs. Smaller cities and industries, especially those lacking developed skills or strong digital infrastructure, will not just fall behind; they will lose top companies, along with the positive effects that lead to high wages and further development. The concept of regional innovation clusters needs to be rethought. Instead of seeing these clusters as universally beneficial and ready for expansion, we must recognize that concentration can occur naturally due to the technology itself. Policies that only spread training and create satellite hubs will fail if they do not also address the factors that attract firms to a few key locations.

Policy Responses: Moving from Readiness to Redistribution and Market Structure

With these realities in mind, if the problem is structural labor redundancy, then public policy must change. First, worker security must be protected and expanded in ways that do not only focus on training. Third, there should be a focus on market structure to reduce dominant firms' ability to capture all the profits generated by AI. Policies on purchasing, open-model standards, data trusts, and public funding for basic computing for public tasks can reduce winner-takes-all trends without stopping private innovation. When the public sector supports core models in areas like health, education, or climate, smaller firms can compete by offering services and local applications, thereby creating more distributed employment. Fourth, regional strategies must be realistic. Instead of trying to create identical major cities, investments should focus on complementary regional strengths, such as specialized clusters that leverage local resources and connect to national infrastructure. This requires more than just subsidies for AI training centers. It requires focused investment in access to computing and broadband, local research collaborations, and governance that allows data and talent to flow in ways that benefit everyone. It is expected that some critics will say that redistribution stifles innovation, while market forces will create new jobs. To some extent, they are correct: innovation does create new jobs. Data on automation and employment show significant disruption focused on certain skills and locations (Ganuthula & Balaraman, 2025). Policy must act to shorten that time. Market structure does not mean total control; it means creating smart rules to shape behavior, such as purchasing practices that reward fair labor, tax assistance for investments that create jobs, and public resources that lower obstacles for small firms. These actions do not prevent growth, but support growth within a social structure.

It is important to stop treating AI readiness as just a simple checklist and start thinking of it as a political and economic issue. The rapid progress of models has a social effect: it concentrates production, wages, and talent in small segments of the workforce and regions. This issue is best called structural labor redundancy. A policy that ignores this risk leads to mass training programs that do not result in jobs and waste public money. It is best to combine worker security, redistribution, and market structure to influence innovation for shared benefits. Invest in public computing resources, regional support, portable benefits, and contracts that encourage labor-positive actions. To allow AI to enable joint success, there must be institutions that recognize the scale of change and act on it. Otherwise, a future is created in which a few highly skilled workers thrive while many communities struggle with job loss, reduced wages, and weakened local capacities.

References

Acemoglu, D. (2018) Artificial Intelligence, Automation and Work. Cambridge, MA: National Bureau of Economic Research.
Baily, M. N., Brynjolfsson, E. & Korinek, A. (2023) Machines of Mind: The Case for an AI-Powered Productivity Boom. Washington, DC: Brookings Institution Press.
Cao, A. (2025) ‘WAIC Shanghai: China reveals new great leap forward with 1,509 AI models’, South China Morning Post.
Demombynes, G. (2025) The Exposure of Workers to Artificial Intelligence in Low- and Middle-Income Countries. Washington, DC: World Bank.
Harris, K., Kimson, A. & Schwedel, A. (n.d.) Labor 2030: The Collision of Demographics, Automation and Inequality. Boston, MA: Bain & Company.
Joshi, S. (2025) The Transformative Impact of Artificial Intelligence on US Labor Markets: Workforce Disruption, Skill Evolution, and the Emergence of Prompt Engineering. Preprints.org.
McKinsey Global Institute (2024) The State of AI in 2024: Generative AI’s Breakout Year. New York: McKinsey & Company.
MERICS (2025) China’s AI Drive Aims for Integration Across Sectors: A Wake-up Call for Europe. Berlin: Mercator Institute for China Studies.
Muro, M. (2026) How the U.S. Can Maintain Its Edge in AI Without Leaving Workers Behind. Washington, DC: Brookings Institution.
OECD (2023) OECD Employment Outlook 2023. Paris: OECD Publishing.
OECD (2026) ‘AI use by individuals surges across the OECD as adoption by firms continues to expand’, OECD Announcement, 1 January.
Restrepo, P. (2025) We Won’t Be Missed: Work and Growth in the AGI World. Cambridge, MA: National Bureau of Economic Research.
State Council of the People’s Republic of China (2025) ‘AI+ initiative and digital economy strategy’, Government Report.
Xu, D., Yang, H., Rizoiu, M. & Xu, G. (2025) ‘From occupations to tasks: A new perspective on automatability prediction using BERT’, arXiv preprint.