By Han Jian and Guo Jinghao

From Silicon Valley to Europe and China, 2025 saw a wave of global layoffs at many high-profile companies, many of which were attributed to artificial intelligence. But is AI really the main culprit? And if so, what human skills are still needed in the rapidly developing AI era?

In this piece, CEIBS Professor of Management Han Jian and Research Fellow Guo Jinghao unpack what the global layoff wave really tells us about AI, work, and the future of workforce strategy.

While the rapid advancement of AI was perhaps the defining business story of 2025, it was often linked to another extremely important narrative: a wave of international layoffs. The layoffs of 2025, however, while clearly global, were not a simple a technology story. In fact, they reflected a complex convergence of the economic pressure and structural change driven by digital and AI transformation.

Data from Challenger, Gray & Christmas indicate that between 50,000 and 100,000 layoffs worldwide were explicitly linked to AI or automation, mostly in the United States with smaller numbers in Europe and Asia. Companies such as Amazon, Microsoft, UPS, Nestlé and Verizon were prominent examples. Some of these cuts were directly tied to AI—especially in customer service, back-office support and routine technical roles—but other major reasons included weak demand, high interest rates, strategic refocusing and business restructuring.

Global layoffs, local dynamics

The United States is at the centre of this cycle. This is not due to broader economic collapse but to a combination of high borrowing costs, profit pressure, a flexible labour market with weaker protections, and restructuring linked to digital and AI change. Last October alone, US companies announced 153,074 job cuts—the worst October in over 20 years—and by late November, total layoffs had reached about 1.17 million, the highest since the pandemic ^[1].

Of these, around 54,000–55,000 were explicitly attributed to AI or automation, mainly affecting customer service, administration, HR and some technical teams. Firms such as Salesforce, Amazon and IBM openly stated that AI replaced parts of their workforce. Still, AI-related layoffs represented only about 4–5%of total US job cuts ^[2]. This shows that while AI is reshaping corporate workforces, the main drivers are still a wider economic slowdown, cost pressure and strategic restructuring. In many cases AI is less the sole cause than a powerful justification for change.

In Europe, many companies reportedly announced layoffs or hiring freezes, especially in technology, manufacturing, automotive, banking, telecoms and equipment. Companies mostly cited cost control, profit pressure, restructuring, and digital efficiency. Automotive and manufacturing firms such as Continental, Bosch and Daimler Truck have been hit particularly hard, along with insurers, renewable-energy companies and multinationals ^[3]. Banks and service sectors also tightened staff. AI and digitalisation are part of these strategies, but usually as elements of broader restructuring rather than the main cause. Economic pressure and sector transformation remain the dominant forces.

China, meanwhile, presents a different picture. The country has not produced nationwide statistics that explicitly attribute layoffs to AI. Broad employment pressure in 2025 has been driven primarily by macroeconomic factors. By last November, unemployment among 16–24-year-olds (excluding students) remained high at 16.9%, while unemployment among 25–29-year-olds was 7.2% ^[4]. Long-cycle struggles in the property markets, partial manufacturing relocation, and subdued consumption are the main drivers of this pressure.

At the firm level, however, technology-related restructuring has become increasingly visible. Specific examples show workforce changes linked not only to economic weakness but also to business portfolio shifts and technological upgrading. Alibaba’s workforce declined sharply between 2022 and 2025 largely due to divestments and organisational restructuring ^[5], while Baidu reduced 20–30% of staff in certain non-core units. In China, AI typically acts as a catalyst embedded within broader restructuring processes rather than a headline reason for layoffs, in contrast to the US where AI is more frequently cited explicitly.

Taken together, evidence from the US, Europe and China shows that the 2025 was of layoffs reflects the interaction of AI-driven transformation with economic stress and broader shifts in corporate strategy. While AI is clearly changing management structures and is increasingly cited as an independent reason for workforce reshaping, especially in white-collar and back-office roles, macroeconomic conditions, rising costs, strategic realignment and efficiency needs remain the main forces. AI plays a dual role: it is both a real driver of increased productivity and a narrative tool that helps companies justify structural change and stricter cost discipline.

AI and the reshaping of corporate workforces

At the corporate level, AI is not merely a technology—it is a capital-intensive project. Turning AI from promise into productivity requires vast and continuing investment in digital infrastructure, computing capacity, data capabilities, model deployment, and security and compliance architecture. These are long-cycle, capital-heavy commitments.

Data from Statista suggests that in 2025 alone, Meta, Alphabet, Amazon and Microsoft are expected to spend hundreds of billions of dollars in capital expenditure, much of it channelled into AI-driven data centres and computing infrastructure ^[6]. Similar commitments are being made by leading Chinese technology companies including ByteDance, which plans to invest roughly RMB 160 billion (around USD 23 billion) in 2025–2026, with about half committed to advanced AI chips such as Nvidia’s H200 and to computing infrastructure ^[7]. IDC forecasts that global AI infrastructure spending will exceed USD 200 billion by 2028, while global cloud infrastructure spending reached USD 102.6 billion in Q3 2025, up 25 percent year-on-year ^[8]. To finance this “capital engineering” of AI, many firms inevitably look to scrutinise labour costs. What follows is not a simple one-to-one substitution of machines for people, but a deeper structural reshaping: companies are reconfiguring both capital and labour—squeezing the middle, decomposing workflows, pushing more activities to platforms and outsourcing.

Recent redundancy announcements show that the roles most exposed are those defined by repetition, clear rules and information-processing intensity: customer support, basic content editing, standardised reporting, data entry, as well as parts of traditional quality assurance maintenance coding and software development. In these domains, AI offers an attractive mix of efficiency, savings, stability and scalability, making them the first targets when firms seek to optimise labour structures and improve returns. This aligns with a recent MIT study estimates that AI could theoretically replace about 11.7% of US jobs, affecting roughly USD 1.2 trillion in wages ^[9]. Because these jobs are highly standardised with clear decision boundaries, once AI and automation tools mature, substitution—or fundamental redesign—becomes much easier.

Swedish fintech company Klarna is a case in point. The company’s AI assistant, developed with OpenAI, handled 2.3 million customer conversations within its first month, around two-thirds of all interactions—equivalent to the work of 700 full-time agents. While Klarna cut about 700 support roles in 2024 as automation ramped up, it partially rehired in 2025 to manage complex and high-value cases, settling on a hybrid model: AI for routine tasks, humans for nuanced judgement and customer trust ^[10].

The lesson is clear: AI is not merely replacing jobs; it is restructuring entire service chains and organisational architectures.

Talent scarcity and the rise of meta-skills

AI is also generating new and more demanding talent needs. AI Workforce Consortium 2025 reports that AI-related job postings are growing faster than the available talent pool, creating structural scarcity^[11]. AI skills are spreading rapidly across functions, with capability expectations being rewritten beyond technical roles into marketing, communications, project management, operations and even executive support. Increasingly, firms are not simply asking whether employees understand AI, but whether they can translate it into real productivity.

Once repetitive work is stripped away, the key question emerges: what capabilities must be kept? In other words, where does true human “irreplaceability” lie in an AI-driven age?

High-risk roles are mirrored by those whose value rises precisely because AI is advancing—roles that are cross-disciplinary, cognitively intensive, and combine capabilites. Broadly, these fall into two groups.

The first consists of technically qualified AI specialists: experts in machine learning, data engineering, MLOps, and cloud and computing infrastructure, as well as AI safety and governance. Demand for such talent is expanding sharply; they are becoming scarce strategic resources that define firms’ competitive positioning.

The second group comprises high-value, human-centred business enablers. They may not write code, but they understand customers, operations and industry context. They know how to embed AI into real workflows, redesign processes, improve service experience and steer organisational change. They are the crucial translators between technical capability and commercial value.

Career platform data suggests that AI skills have already featured in around 45% of executive job requirements and are spreading rapidly across finance, operations, design, sales and other non-technical positions—evidence that AI fluency is becoming a core competency in cross-functional roles ^[12].

As AI penetrates more sectors, soft skills and higher-order cognitive capabilities become even more critical. Communication, coordination, leadership and cross-department collaboration remain difficult for AI to replicate, and are increasingly decisive for organisations deciding whom to retain and redeploy. Demand is also rising for hires that are able to interpret model bias, manage ethical and compliance risk, and bring a governance mindset to AI deployment.

Surveys by McKinsey, BCG and PwC show that future resilience will not rest on mastery of any single tool but on a portfolio of meta skills—an aptitude for continuous learning, adaptability, deep business understanding, and collaborative problem-solving ^[13][14][15]. These meta-skills not only help individuals withstand role restructuring. They also underpin corporate returns on AI investment and the release of real productivity.

AI is therefore not merely changing which tasks are performed; it is reshaping how value, productivity, and competitiveness are defined.

Companies must now rethink three questions: which work must remain human; which can be automated; and which capabilities must be retrained and upgraded? Talent is no longer just a cost—it is the decisive strategic lever that determines whether enterprises can truly unlock AI’s potential and sustain competitiveness. This is not a short-term disruption, but a multi-year process of deep restructuring.

Leadership and the rewriting of the AI narrative

From the perspective of management education and leadership development, workforce transformation in the AI era is not a matter of upgrading tools. It is a systemic test of organisational capability, leadership philosophy and corporate purpose.

Recent corporate experience has repeatedly shown that if leaders treat AI merely as a cost-cutting instrument, the result is often superficial efficiency gains rather than sustainable productivity or innovation. Business schools must therefore place “leading AI transformation” at the centre of their agenda—integrating technological understanding, organisational redesign and talent strategy to cultivate leaders capable of building long-term value.

Effective AI leadership requires reframing the narrative, from “replacing people” to “reconstructing work and value creation”. This narrative shapes trust, psychological security and an organisation’s willingness to learn about, experiment with, and adopt AI.

Equally important is long-term investment in talent. Reskilling cannot remain rhetorical; it must become operational strategy. Leading firms demonstrate that downsizing and reskilling are not mutually exclusive.

Although companies such as IBM, AT&T, Siemens, Accenture, Amazon and Microsoft have all conducted layoffs over the past year, they have also invested heavily in employee reskilling and capability upgrading. Their shared lesson is that long-term competitiveness is built not by cutting alone, but by continuously upgrading their human capital as well.

Leaders must also acquire the capability to redesign processes and restructure organisations. BCG warns that simply layering AI onto legacy workflows delivers only fragmented benefits. Only by reshaping end-to-end processes, and allowing humans and models to each focus on what they do best, can true productivity gains emerge^[16].

In parallel, business schools should enhance education on psychological safety, learning cultures and ethical governance. PwC’s research shows that permission to experiment, active learning support and avoidance of fear-based management are critical to sustainable AI adoption ^[17].

Corporate thinking about “people efficiency” must be rewritten. If companies treat productivity gains merely as an excuse to remove a few more workers, then their ambitions are too narrow. A large body of economic and management scholarship supports the view that technological change need not turn surplus labour into “waste,” but can instead create new productive capacity when organisations choose to invest in people.

From a management perspective, high-performing firms succeed not by eliminating labour, but by continuously reconfiguring skills and redeploying people into higher-value roles ^[18].

Historical evidence—from the Industrial Revolution to the IT era—confirms that economies and firms that invest in skill upgrading and organisational redesign at times of technological disruption achieve greater innovation, stronger resilience, and more sustainable growth than those that rely solely on shedding labour.

Beyond the headlines

The global layoff narrative, therefore, should not focus only on short-term and attention-grabbing headlines about AI but be viewed within the larger context of a long cycle of talent strategy and organisational evolution.

AI is not reshaping one or two occupations; it is rewiring the logic of workforce strategy itself. Companies must continually reassess which work must remain in human hands, which can be automated, and which capabilities must be retrained and upgraded.

This should also present a lesson to managers and firms. Ultimately, truly strong companies do not rely solely on layoffs to cut costs—they use reskilling to turn people into drivers of future growth.

Han Jian is Professor of Management and Director of CEIBS Centre for Organisational Growth and Talent Development at CEIBS. Guo Jinghao is a Research Fellow at CEIBS.