AI Won’t Replace Most Workers — But It May Replace Their Leverage

By Kael Rosenberg | Updated on May, 2026 | 🕓 12 minutes

Key Highlights

- How AI is reshaping leverage in the workplace, rather than eliminating jobs.

- The three layers of work leverage: skill, information, and scale.

- Practical steps for knowledge workers to protect and rebuild leverage in the AI era.

If you closely observe the global labor market from 2023 to 2025, you will notice a strange phenomenon: large-scale unemployment hasn’t occurred, yet more and more people feel that their job remains, but their sense of value is eroding.

The International Labour Organization (ILO), in its World Employment and Social Outlook: Trends 2024, reported that the global unemployment rate in 2024 remained at 4.9%, even lower than pre-pandemic levels. PwC’s 2025 Global AI Jobs Barometer, analyzing nearly one billion job advertisements across six continents, found that in the industries most exposed to AI, the number of positions actually grew by 38%. The data looks optimistic.

Yet another set of data reveals cracks. The ILO’s Global Wage Report 2024-25, tracking data from 1999 to 2024, found that in high-income countries, labor productivity grew by 29%, while real wages only increased by 15%. More crucially, the report noted that the lowest-earning 10% of workers globally receive only 0.5% of total wages, while the highest-earning 10% take nearly 38%.

Thus, the question is not “Will AI make me unemployed?” Instead, the question is: AI is systematically reshaping the “leverage structure” of work—not eliminating jobs, but weakening most workers’ bargaining power within their positions.

What is “Work Leverage,” and Why Is It More Important Than the Job Itself?

I like to divide “work leverage” into three layers. This is not an academic theory but a practical framework based on observing different industries:

The first layer is skill leverage. You command a scarce skill that the market is willing to pay a premium for. An engineer who can write complex embedded systems and a front-end developer who only knows basic web development have very different income ceilings.

The second layer is information leverage. You possess tacit knowledge about a field that your client doesn’t. Doctors know the real risk distribution behind certain symptoms; lawyers know the enforceability of clauses in case law. This information asymmetry forms the pricing foundation of professional services.

The third layer is scale leverage. Your output-to-income relationship is not linear. A top surgeon can, by leading a team and establishing processes, generate ten times the revenue of a solo practitioner. This is the leap from “selling time” to “selling a system.”

AI’s impact on these three layers of leverage is far more complex than a simple binary narrative of “replace or not replace.”

First Layer Erosion: Skill Leverage and the “Collapse of the Middle Tier”

The most visible impact of AI is the compression of the skill value distribution. It lowers entry barriers to near zero while sharply reducing the value of “mid-level proficiency.”

Take the translation industry as an example. A study titled Lost in Translation: Artificial Intelligence and the Demand for Foreign Language Skills tracked translation employment and Google Translate search volume, finding a significant correlation. This does not mean translation work has disappeared—literary translation and cross-cultural negotiations still require humans—but the mid-tier skill of “accurately translating a passage from English to Chinese” has shifted from a scarce asset to a replaceable cost. In Japan and South Korea, this has been particularly noticeable: budgets for technical documents and basic content translation have been slashed by 60% to 80%, turning translators from “creators” into “AI output quality controllers.”

The U.S. customer service industry provides another perspective. According to Site Selection Group, between 2022 and 2024, around 80,000 U.S. customer service positions were eliminated. Yet many companies did not completely remove human agents; instead, they set AI resolution rates as the new KPI benchmark. Human agents now handle the “leftovers” that AI cannot solve—workload has increased, but salaries have not, because employers redefined the positions as “low-cost AI-assisted support roles.”

Changes in India’s IT outsourcing sector are even more cautionary. Wipro and Infosys slowed hiring of junior programmers in 2024 while adding positions for “AI orchestrators.” MIT Sloan’s paper The EPOCH of AI: Human-Machine Complementarities at Work (Loaiza & Rigobon, 2024) explores this human-AI complementarity: AI does increase output for some senior developers, but the middle tier—those “better than junior but not yet architects”—see their skill premiums rapidly evaporate. They are not fired, but promotion pathways narrow, because AI fills the experience gap they would have spent three years accumulating.

PwC’s 2025 Global AI Jobs Barometer also highlights an easily overlooked point: in AI-exposed industries, the pace of skill demand changes has accelerated by 66%. This means the era of “master one skill over ten years” is ending.

Second Layer Erosion: Information Leverage and the “Transparency Trap”

If skill leverage erosion is visible, the collapse of information leverage is more subtle—it occurs in the power dynamics between you and your client.

Europe’s legal tech market is undergoing this transformation. Mid-sized law firms in Germany and the U.K. that adopted AI contract review tools have encountered a paradox: client satisfaction did not drop, but the average fee per client did. Why? Because clients also started using AI. When corporate legal teams first run contracts through ChatGPT before consulting a lawyer, the lawyer’s information monopoly is broken. Lawyers shift from “knowledge authorities” to “verification nodes,” which naturally command lower fees than diagnostic nodes.

This “transparency trap” also appears in financial planning. In Brazil and Mexico, the rise of digital banking apps allows ordinary middle-class individuals to generate basic investment portfolios on their own. The World Bank’s Women, Jobs, and the Impact of Artificial Intelligence (2024) specifically notes the dual effect on women: while it lowers entry barriers, it also compresses incomes for mid- to low-level financial advisory roles traditionally dominated by women.

The collapse of information leverage does not mean professional expertise is worthless. Complex mergers or rare disease diagnoses still require deep specialists. But the once-protective moat of basic information asymmetry, which supported the incomes of many middle-class professionals, is drying up quickly.

Third Layer Erosion: Scale Leverage and the “Transfer of Ownership”

This is the deepest layer and the least discussed.

Traditionally, individuals could amplify output through accumulated experience and methodology. A senior consultant could handle five projects by mentoring junior consultants, leveraging their time. Now, AI has become a new amplifier—but the problem is that this amplifier usually belongs to the company or platform, not the individual.

In content creation and creative industries, this transfer is particularly evident. A senior graphic designer once relied on “aesthetic judgment + software proficiency + client understanding” to establish scale leverage. Now, when a company purchases enterprise versions of Midjourney, Stable Diffusion workflows, and an internally trained style LoRA, the “style” itself becomes encoded into the company’s digital assets. The designer’s personal brand is diluted, and they become just one operator of the company’s AI assets.

Another subtle manifestation of scale leverage transfer is that as AI allows “senior experience” to be partially encoded and replicated, the irreplaceability of senior staff decreases. Companies no longer need as many senior experts to maintain output, because a combination of AI, a few senior staff, and many junior operators can replace the traditional pyramid team structure.

Acemoglu (2024), modeling GenAI’s effects on wages and inequality, makes a key observation: in almost every scenario, AI deployment may exacerbate inequality between capital and labor. The ILO’s working paper in Latin America (Exposure to Generative AI and the Digital Divide in Latin America and the Caribbean, 2024) cites this analysis, noting that although AI adoption remains low in the region, labor markets already exhibit “skill-biased technological change”—high-skill workers’ gains and low-skill workers’ struggles are widening simultaneously.

Self-Check: Is Your Leverage Eroding?

Instead of anxiously wondering if AI will replace you, it’s more practical to evaluate your vulnerability. Here are five high-vulnerability signals I’ve observed—if you match two or more, it’s time to rethink your positioning:

1. Your core output can be clearly described as a linear “input → process → output” workflow. AI excels at optimizing such workflows.

2. Your client or supervisor tries AI solutions before reaching out to you. Your role is shifting from “first-choice solution” to “backup verification.”

3. Your billing is still primarily time- or piece-based rather than outcome- or risk-based. As AI compresses output per unit time, time-based billing becomes self-devaluing.

4. Your industry has seen low-cost, AI-assisted competitors in the past three years. This shows that the market is already redefining pricing benchmarks using AI.

5. Your skill updates rely mainly on “learning new tools,” not “building new networks or cognitive frameworks.” Tool-level catch-up cannot keep pace with AI iteration speed.

Conversely, low-vulnerability traits typically include: work involving multi-stakeholder coordination and tacit political judgment; accountability for results legally or reputationally; value built on long-term trust accumulation and deep contextual understanding.

Rebuilding Leverage: Four Action Directions in Uncertainty

Acknowledging uncertainty is the first step. No one knows exactly what the labor market will look like in five years. IMF’s Gen-AI: Artificial Intelligence and the Future of Work (2024) and concurrent World Bank studies emphasize this uncertainty. But uncertainty does not equal inaction. Here are four practical action directions at this stage:

Direction One: From “Skill Holder” to “Context Integrator”

Don’t compete with AI by “doing things faster”—compete by “knowing when and what to do.” Build cross-domain connection capabilities. A programmer who only knows code is vulnerable, but a programmer who knows code + a vertical industry (e.g., supply chain, healthcare compliance) + business logic possesses AI-resistant “contextual judgment.” AI optimizes within known frameworks, but humans excel at judging “whether the problem is being asked correctly.”

Practical suggestion: Spend five to ten hours each quarter deeply understanding an adjacent field. Not just shallow reading, but talking to practitioners to understand real pain points and decision logic.

Direction Two: From “Information Monopoly” to “Trust Asset”

When information is no longer scarce, being trusted becomes the new scarce resource. Actively showcase your decision-making process and failure cases. In a world where AI can generate perfect answers, transparency becomes a differentiator. Clients are willing to pay for “I know this advisor will tell me the risks AI won’t.”

Practical suggestion: Organize your decision logic and lessons learned into an internal knowledge base or public sharing (within industry compliance limits). This is not revealing trade secrets—it’s building a “trusted decision-maker” brand.

Direction Three: From “Time-Based Billing” to “Outcome/Risk-Based Billing”

Change your pricing unit so that part of AI efficiency gains translates into your profit rather than the employer’s cost savings. Repackage your service as a “results commitment.” Don’t sell “I’ll work 80 hours”; sell “I’ll help you pass the audit.” Introduce a base fee + success fee model to lock in part of AI-driven efficiency gains.

Practical suggestion: Choose your most confident client and try shifting the next project from hourly pricing to results-based pricing. The first attempt may be difficult, but it forces you to redefine your value proposition.

Direction Four: From “Individual Output” to “System Ownership”

Become the architect of AI workflows rather than just an operator. Build AI workflows, knowledge bases, and judgment frameworks exclusive to your industry or client base. This is not teaching others to use ChatGPT—it is designing a “human+AI collaboration system” only you can fine-tune. Future managers’ core value lies in designing “human+AI” collaboration topologies, not just managing people.

Practical suggestion: Document your complete AI workflow—what steps require human intervention, and why. These judgment criteria are your intellectual property. Organize them into a transferable framework, even if for internal use only.

Conclusion: Survival Rules in the Era of Leverage Transfer

AI will not eliminate jobs, but it will redistribute power within them. The key in the next few years is not “learning AI”—AI tools’ learning curves are flattening monthly—but repositioning yourself to an irreplaceable spot in the value chain.

The ILO’s Global Wage Report 2024-25 reminds us that the long-term decline in global labor share has been ongoing for decades, and AI only accelerates this structural tension. PwC’s optimistic data and ILO’s cautionary data are not contradictory—they describe two sides of the same coin: job numbers may grow, but bargaining power within positions is shifting.

Thus, the core question for every knowledge worker is: If your client could complete 80% of your work using AI tomorrow, how much would they pay for your remaining 20%?

How much that 20% is worth depends on how you start building it today.

FAQs

Q1: If AI doesn’t eliminate most jobs, why do people feel less valued at work?

A1: AI changes *how much leverage workers have - rather than simply removing jobs. Tasks that were once rare or required specialized judgment can now be partially automated, reducing bargaining power. Even if your role remains, the value you can claim in negotiations, promotions, or project ownership may be shrinking.

Q2: Which types of jobs are most at risk of leverage erosion due to AI?

A2: Mid-level or routine-skilled roles are most exposed—jobs that are neither entry-level nor highly specialized. For instance, basic translation, standard legal document review, or repetitive IT coding tasks are increasingly supported by AI, which reduces their premium in the labor market. Senior experts or roles that require deep judgment, cross-domain integration, or trust remain less vulnerable.

Q3: Can AI actually create new opportunities for workers?

A3: Yes. AI can amplify productivity for those who know how to integrate it effectively. Roles like AI workflow architects, human-AI collaboration managers, or specialists who leverage AI insights to make better decisions are emerging. The key is to shift from “doing tasks faster” to “deciding what to do, when, and why.”

Q4: What practical steps can professionals take to maintain or rebuild leverage in an AI era?

A4: Focus on aspects AI cannot easily replicate:

- Develop cross-domain expertise to integrate knowledge from multiple areas.

- Build trust and transparency with clients, showing judgment and accountability.

- Shift pricing models from hourly work to outcome- or risk-based approaches.

- Take ownership of AI workflows and intellectual property within your domain, becoming the “architect” rather than just the operator.

References

1. PwC. 2025. Global AI Jobs Barometer. London: PwC. Analysis based on close to one billion job advertisements from six continents.

2. International Labour Organization (ILO). 2024. Global Wage Report 2024-25: Is Wage Inequality Decreasing Globally? Geneva: ILO.

3. International Labour Organization (ILO). 2024. World Employment and Social Outlook: Trends 2024. Geneva: ILO.

4. Gmyrek, P., et al. 2025. Generative AI and Jobs. International Labour Organization Working Paper, May 2025.

5. World Bank. 2024. Women, Jobs, and the Impact of Artificial Intelligence. Washington, DC: World Bank.

6. International Monetary Fund (IMF). 2024. Gen-AI: Artificial Intelligence and the Future of Work. Washington, DC: IMF.

7. Loaiza, I., and Rigobon, R. 2024. "The EPOCH of AI: Human-Machine Complementarities at Work." MIT Sloan Research Paper No. 7236-24, November 21, 2024.

8. Acemoglu, D. 2024. Modeling of GenAI outcomes on wages and inequality. Referenced in ILO Working Paper 121, 2024.

9. International Labour Organization (ILO). 2024. Exposure to Generative AI and the Digital Divide in Latin America and the Caribbean. ILO Working Paper 121.

10. "Lost in Translation: Artificial Intelligence and the Demand for Foreign Language Skills." Cited in AI Multiple Industry Analysis, 2025.

11. Site Selection Group. 2024. Customer service employment data, United States, 2022–2024.

About the Author

Kael Rosenberg, MBA is a technology and labor market analyst focusing on how AI reshapes work, productivity systems, and creative economies. He holds an MBA from London Business School and has worked as a consultant for digital transformation projects in Fortune 500 companies. His research explores how AI changes labor leverage, creative ownership, skill hierarchies, and the evolving definition of “knowledge work” in the automation era.

Editorial Transparency Statement

This article is based on publicly available data, research papers, and industry reports. All sources are cited in the references section to ensure accuracy, traceability, and transparency.

The analysis and conclusions are the author’s own interpretation of the data. No external influence affected the content or recommendations.

Disclaimer

This article is for informational purposes only and does not constitute professional advice. Readers should conduct their own research or consult qualified experts before making decisions related to career, finance, or business strategies.