Adecco Study Finds AI Will Transform Jobs Without Causing an Employment Collapse

The last few years have brought generative AI from research labs into everyday tools, from coding assistants to customer-service chatbots and recruiting agents. That shift has intensified long‑standing fears of a “jobless future,” but the empirical evidence still points to something more nuanced: AI is becoming a powerful engine for job redesign, productivity, and task reshaping rather than mass unemployment. Recent global estimates suggest that one in four jobs worldwide is now in occupations with some degree of exposure to generative AI, capturing the scale of this shift. Furthermore, the opacity of algorithms can exacerbate the risks of bias in AI hiring processes.

Generative AI is emerging as an engine of job redesign and productivity, not widespread unemployment.

Institutions that track labour markets over time—OECD, the IMF, the World Economic Forum and others—converge on a similar view. They find significant exposure of tasks and occupations to automation, widespread anxiety among workers, and real displacement in specific roles, but limited aggregate employment effects so far.

From Automation Fears to Task-Level Transformation

Concern about machines replacing human labour is not new. Early automation waves in manufacturing and clerical work altered routine tasks but did not eradicate employment; instead, they shifted workers into new roles and industries. AI is extending this pattern from physical automation to cognitive work, with generative systems now able to handle text, code, and increasingly multimodal tasks.

OECD analysis of jobs that were previously flagged as “high risk of automation” shows that these occupations did not disappear; they continued to grow, albeit more slowly than low‑risk jobs. This is a crucial historical lesson: forecasts based on technical feasibility often overstate how quickly employers reconfigure work and downplay the creation of new, complementary tasks.

At the same time, adoption of advanced AI remains relatively low across firms, concentrated in larger companies that are still experimenting with these technologies. That slow diffusion helps explain why, despite intense hype, there is “little evidence of significant employment effects” from AI so far in company-level data. The technology is powerful, but organisations adapt more gradually than headlines suggest.

What the Data Actually Say About Jobs and AI

Several strands of evidence now give a clearer picture of how AI is affecting work:

• When AI-driven automation is factored in, occupations at highest risk of automation account for roughly 27% of employment in OECD countries. That is a substantial share, but still far from “most jobs.”
• Over 2012–2019, employment grew in nearly all occupations analysed, and there is no clear relationship between AI exposure and employment growth overall.
• In fact, in occupations where computer use is high, greater exposure to AI is associated with higher employment growth, suggesting a complementary effect rather than simple substitution.
• Across multiple surveys of employers in manufacturing and finance, most report no change in overall employment due to AI so far, with some even reporting increases.
• New case studies across eight OECD countries show that job reorganisation is more prevalent than job displacement, as tasks are reoriented toward what humans do better—judgment, interaction, and complex problem-solving.

Layered on top of this OECD work, IMF analysis estimates that around 40% of global employment is affected in some way by AI—through automation risk, augmentation, or changing skill requirements—with advanced economies more exposed but also better positioned to capture productivity gains. That dual reality—high exposure and high potential upside—defines the current phase.

Hybrid Workflows in Practice: Adecco’s Recruitment Example

Adecco’s recent whitepaper offers a concrete case study of what this transformation looks like inside a major employer. In its recruitment operations, the company has deployed AI agents across seven stages of the hiring lifecycle, turning recruitment into a continuous, hybrid human–AI process rather than a linear sequence of manual steps.

Instead of replacing recruiters, the system integrates automation with human judgment: AI agents handle screening, scheduling, and routine interactions, while people focus on complex candidate evaluation, relationship building, and nuanced client decisions. According to Adecco, this redesign has:

• Cut time‑to‑deliver by roughly 50 percent
• Pushed fill rates above 80 percent
• Generated about 1.2 million AI‑powered candidate–agent interactions
• Completed around 250,000 interviews across 50,000 roles

Adecco frames this as “hybrid workforce orchestration”—a deliberate combination of people, AI agents, software automation, and physical systems to maximise value creation rather than remove humans from the equation. That framing matters: it treats AI not as a standalone replacement technology but as one component in a broader work system.

Net Job Creation vs Displacement: The Numbers Behind the Narrative

The Adecco–Altermind analysis positions AI as a general‑purpose technology that ultimately expands employment and raises real wages, even as it displaces some roles in the short term. In one workforce study cited in the report, AI contributed to the displacement of about 3.7 million jobs globally between 2019 and 2024 while supporting the creation of roughly 6 million new AI‑related positions, implying a net gain of 2.3 million roles.

It further projects a similar net job creation potential—again around 2.3 million—by 2030, while explicitly highlighting the uncertainty around those forecasts.

Those estimates sit within a wider range of global scenarios. A synthesis of institutional data suggests that by 2030 AI could displace around 92 million roles while creating approximately 170 million new ones, for a net gain of about 78 million jobs globally. Other analyses forecast a net gain of 15–25 million jobs, with new roles concentrated in technology, healthcare, and services while routine cognitive jobs shrink.

The precise numbers differ, but the direction is remarkably consistent across sources: AI drives substantial job churn—some tasks and roles shrink, others grow—but the net effect is more creation than destruction. That pattern reflects what labour economists call the “reinstatement effect,” where productivity gains and new capabilities generate additional tasks, occupations, and demand for workers whose skills complement intelligent systems.

Importantly, these projections carry non-trivial uncertainty. They depend on policy choices, investment levels, regulation of AI, and how quickly organisations reengineer workflows. Trustworthy analysis needs to treat them as directional signals, not precise forecasts.

Uneven Impacts Across Sectors and Skills

The impact of AI is not evenly spread. Sectoral and occupational data highlight clear patterns:

• Routine, codifiable knowledge work—data entry, basic customer support, standardised reporting—is highly exposed because many of its tasks can be specified and automated or heavily augmented by AI.
• Manual, interpersonal, and care-intensive roles—nursing, elder care, hospitality, skilled trades—remain far less exposed to direct automation, though they may still be affected indirectly through scheduling, resource allocation, and digital tools.
• In workplaces most exposed to AI, demand for some traditional business, management and digital skills has shifted, with evidence of both restructuring of roles and changes in the specific skill mix employers seek.
• Across 10 OECD countries, job postings indicate rising demand over time for emotional, social, and digital skills—roughly 15% increases—as well as about an 8% rise in business and management skills and significant growth in cognitive and language skills, especially in occupations highly exposed to AI.

These skill shifts reinforce a simple message: AI increases the value of being able to work with technology and with people. Technical proficiency alone is not enough; the roles that grow tend to combine digital capabilities with judgment, communication, and collaboration.

What This Means for Businesses

For employers, the evidence points to a clear strategic agenda:

1. Treat AI as a work-design problem, not just a technology deployment. Case studies show that the biggest gains come when organisations redesign tasks and workflows—rather than bolt AI onto existing processes—so humans focus on higher-value work.
2. Invest systematically in reskilling and upskilling. OECD surveys link training to better outcomes for workers and to higher trust in how AI is implemented. Building AI literacy, data skills, and human-centric capabilities is now a core business function, not a side project.
3. Co-create AI adoption with workers. Consultation and transparency around where and how AI is deployed correlate with better job quality and stronger trust. Involving employees in redesigning their own roles can surface risks early and unlock practical ideas for augmentation.
4. Measure job quality, not just productivity. Workers in many AI-enabled settings report improvements in enjoyment and safety—thanks to automation of tedious or dangerous tasks—but also higher work intensity and new stressors. Responsible leaders track both and adjust.
5. Build governance around bias, privacy, and agency. Leading policy work highlights risks of discrimination, loss of control, and lack of transparency if AI is deployed without safeguards. Clear accountability, auditability, and human override mechanisms are essential.

Implications for Workers and Society

For workers, the data suggest that the most resilient careers will be built around adaptability and complementary skills rather than static job titles. Emotional intelligence, communication, and complex problem-solving are rising in importance alongside digital competencies. That combination positions people to work with AI systems instead of competing directly against them.

At the societal level, the key risk is not mass unemployment in rich economies; it is uneven adjustment. Regions, sectors, and workers with less access to training, digital infrastructure, and safety nets may experience the churn of displacement without seeing the benefits of new job creation. Policies that support lifelong learning, portable benefits, and targeted transition assistance will heavily influence whether AI-era labour markets are inclusive or polarised.

There is also a trust dimension. Surveys show many workers are worried about job loss and wage impacts from AI, even when their employers report limited changes so far. Bridging this perception gap requires honest communication, visible worker participation in AI deployment, and sharing productivity gains through better wages, job quality, or reduced precarity.

Looking Ahead: AI as a Catalyst for Job Redesign

Taken together, the evidence depicts AI as a catalyst for job redesign rather than outright destruction. Advanced systems are exposing roughly a quarter to two‑fifths of global employment to task-level transformation, with relatively modest aggregate employment effects so far and a plausible path toward net job creation under most scenarios.

For organisations, the Adecco experience highlights a pragmatic path: orchestrate hybrid human–AI workflows, measure both productivity and job quality, and build processes in which automation amplifies human strengths. For workers, the priority is to cultivate skills—technical, social, and cognitive—that travel well across evolving roles.

The next decade will test whether businesses and policymakers can manage this transition in a way that shares the gains broadly. If they succeed, AI will be remembered less as a job-destroyer and more as a force that reshaped work—unlocking new kinds of roles, new forms of collaboration, and, potentially, higher real wages for those whose skills complement intelligent systems.

Conclusion

The headline takeaway from Adecco’s new study is deceptively simple: artificial intelligence is fundamentally reshaping work, but it is not triggering an employment collapse. That message matters right now because the public conversation is still split between “job apocalypse” narratives and boosterish claims that AI will only create new opportunities. Adecco’s findings fall in neither extreme. Instead, they paint a picture of hybrid labor markets in which tasks, roles, and skills are being reconfigured at speed—and where the outcome for workers will depend heavily on how leaders manage reskilling, regulation, and trust.

From “job apocalypse” fears to a more nuanced reality

Concerns that AI would decimate jobs peaked when generative models first hit mainstream awareness in 2023–2024, prompting waves of predictions that large portions of the workforce could become obsolete within a decade. A 2024 Adecco survey, for example, found that around 41% of senior executives expected their organizations to employ fewer people over the next five years because of AI, reflecting genuine anxiety about automation and cost cutting.

Fast-forward to 2026, and the new Adecco study points to a more complicated reality: AI is driving a “massive evolution in the world of work,” but “a job apocalypse is not on the horizon.” CEO Denis Machuel emphasizes that the technology is “more about changing roles and tasks than eliminating jobs,” underscoring that the primary impact so far is task reallocation rather than widespread job destruction.

This is consistent with broader labor-market data. S&P Global’s 2026 assessment of AI and employment finds that the net impact on jobs has turned modestly negative—about -5 percentage points globally over the past year, with a further -2 points expected—but still far from the kind of structural collapse some feared. The report cautions against interpreting this softening as evidence of large-scale AI-driven displacement, highlighting practical constraints like reliability, security, and the need for human oversight.

The World Economic Forum’s projections add important long-term context: by 2030, it expects job disruption to affect roughly 22% of all jobs worldwide, with around 170 million new roles created and 92 million displaced—a net gain of about 78 million positions. The fastest growth is anticipated in technology, data, AI, healthcare, education, and green economy roles, reinforcing the idea that the core challenge is not job quantity, but job quality, access, and readiness for new skill demands.

What Adecco’s latest study actually says

Adecco’s whitepaper and related analysis center on the idea of “hybrid labor markets,” where humans, AI agents, and physical automation are increasingly intertwined within workflows. Several findings stand out:

1. AI is changing tasks faster than it is eliminating jobs. Multiple sources in the whitepaper show that AI is reshaping tasks, workflows, and skill requirements across sectors, but not yet wiping out whole occupations at scale. In the U.S., the employment share of AI-exposed jobs has remained broadly stable so far, which aligns with evidence from the Bank for International Settlements that firms adopting AI see productivity gains without significant short-term labor replacement.
2. New AI-related roles are emerging. Adecco estimates that between 2022 and 2025, around 1.9 million new AI-related jobs were created, covering areas such as AI product management, model governance, prompt engineering, and human-in-the-loop oversight roles. These roles typically sit at the intersection of business understanding and technical fluency, foreshadowing a broader workforce shift toward “AI generalists” who can direct, interpret, and quality-check AI outputs.
3. Impacts are highly uneven by sector and job type. Entry-level and routine office roles—especially administrative and basic knowledge work—are among the most exposed, with some positions already disappearing or being redesigned. Machuel notes that companies cannot simply cut junior roles without damaging their future talent pipelines, so many are experimenting with reinventing these jobs to integrate AI as a complementary tool rather than a replacement. In contrast, care work, many service jobs, skilled trades, and complex physical-world roles remain relatively insulated for now by real-world variability, interpersonal demands, and practical constraints on automation.
4. Hybrid human–AI workflows are pulling ahead. Organizations that treat AI as a fully integrated element of job design, workflow architecture, and performance models—rather than a bolt-on productivity tool—are already outpacing those that remain in pilot mode. Adecco finds that firms orchestrating people and AI agents together are seeing more consistent productivity gains and innovation capacity than those relying on narrow automation.

Taken together, these points support Adecco’s central conclusion: AI is acting as a powerful force for restructuring work, but the data so far provide no evidence of an impending employment collapse.

How work is being redesigned: the rise of hybrid human–AI roles

Across industries, the conversation has shifted from “Will AI take jobs?” to “How are jobs changing?” Several converging trends define this transition:

• AI generalists as a core workforce profile. PwC’s analysis of the AI-enabled workforce argues that by 2026, organizations will rely heavily on “AI generalists”—professionals who combine broad business understanding with the ability to direct, interpret, and audit AI systems. As agentic AI handles more specialized tasks, human work increasingly centers on judgment, creativity, ethics, and context-sensitive decision-making.
• Connected intelligence and agentic AI. Cisco highlights “Connected Intelligence” as a defining workplace trend for 2026: networks that link people to people, people to AI, and AI to AI. This model depends on agentic AI systems that can collaborate, hand off tasks, and coordinate workflows, transforming roles from executing routine work to overseeing and orchestrating intelligent agents.
• New orchestration and oversight roles. Adecco’s whitepaper notes that AI is not only automating tasks, but creating new categories of work focused on coordination, integration, and governance—functions like AI operations managers, AI ethicists, and system orchestrators who design and supervise hybrid workflows. These roles sit at the heart of the emerging hybrid labor market, acting as a bridge between technical capability and business strategy.
• HR and workforce management transformation. SHRM’s 2026 “State of AI in HR” report finds that AI adoption is driving shifts in job responsibilities and accelerating upskilling, but with relatively modest displacement so far. HR professionals at AI-deploying organizations report slight job displacement (7%), some new jobs or roles (24%), changes in job responsibilities for 39% of workers, and frequent upskilling or reskilling opportunities for 57% of employees.

These trends underscore a critical point: AI is not simply an automation technology; it is a re-architecture technology. The organizations that benefit most will be those that design roles, career paths, and performance expectations around hybrid human–AI systems rather than trying to bolt AI onto old job structures.

The reskilling imperative and shifting skills landscape

Adecco’s broader Workforce Trends 2026 report shows just how quickly AI has climbed the agenda for both workers and executives. Workers now rank AI and generative AI among the top three megatrends reshaping their organizations, up from seventh and ninth place respectively in 2024. C‑suite leaders have kept AI and GenAI in their own top three influences for two consecutive years, signaling that this is no longer a side experiment but a central strategic concern.

This growing prominence is driving a decisive shift in skill strategy:

• From degrees to demonstrable skills. PwC anticipates that as skills evolve at unprecedented speed, employers will move further away from degree-based hiring and toward competency-first frameworks. Skill passports, AI-enabled learning pathways, and granular skill taxonomies are expected to become more common, allowing workers to signal capabilities in AI literacy, data reasoning, and digital collaboration beyond traditional credentials.
• Widespread upskilling and reskilling. SHRM’s data illustrates this change at the organizational level: more than half of HR professionals in AI-adopting firms report frequent upskilling or reskilling opportunities for employees, tied directly to AI implementation. Adecco’s commentary reinforces that reskilling and career transitions will be especially critical for entry-level and routine knowledge workers, whose tasks are most exposed to automation yet who remain essential to long-term talent pipelines.
• Structural job churn, not simple job loss. The World Economic Forum’s forecast of 170 million new roles and 92 million displaced by 2030 highlights the scale of churn that AI and other disruptions are likely to generate. The fundamental risk is that workers who lack access to effective reskilling pathways could find themselves stranded, even as overall employment remains stable or grows.

In practice, this means that “no employment collapse” is not the same as “no problem.” Adecco’s study implicitly frames reskilling, thoughtful regulation, and proactive workforce planning as the determinants of whether AI’s impact will be broadly inclusive, or whether it will deepen inequality between those who can adapt and those who cannot.

Risks, gaps, and uneven impacts

A credible analysis has to account for the downside risks alongside the opportunities. Several warning signs in the current data deserve attention:

• A widening gap between AI ambition and workforce readiness. Adecco’s global study of 2,000 C‑suite executives across 13 countries finds that 45% of business leaders expect AI agents to be integrated into workflows within the next 12 months. Yet only 36% say their talent strategy clearly demonstrates that AI will create opportunities for employees, indicating a growing disconnect between adoption plans and workforce narratives.
• Misalignment between leaders and workers. The same study shows that while 70% of workers feel ready to collaborate with AI agents, only 39% of leaders believe employees would be comfortable doing so. This mismatch can erode trust and slow adoption if workers perceive AI as a threat and leaders underestimate their willingness—or vice versa.
• Modestly negative short-term employment effects. S&P Global’s analysis quantifies a net negative impact on employment from AI investments in the past 12 months, even as longer-term projections remain more balanced. Smaller firms continue to forecast a net positive employment effect (+3 points), but larger organizations show greater short-term job rationalization as they deploy AI.
• Pressure on middle management and routine roles. Gloat’s review of AI workforce trends highlights Gartner’s prediction that through 2026, about 20% of organizations will use AI to flatten their structures, potentially eliminating more than half of current middle management positions. Combined with Adecco’s identification of entry-level office jobs as particularly exposed, this points to potential pinch points at both ends of the career ladder.
• Residual displacement and stress inside organizations. Even in relatively positive HR cases, SHRM finds that AI has led to some job loss: 7% of HR professionals in AI-deploying organizations report job displacement, which, while not catastrophic, represents real disruption for affected workers.

These risks do not contradict Adecco’s finding that AI is unlikely to cause an employment collapse; they refine it. The emerging picture is one of moderate, uneven pressure on employment, with significant local pain points—especially in routine, entry-level, and some managerial roles—counterbalanced by new opportunities for those positioned to move into AI-enabled functions.

Policy and leadership priorities: making “no collapse” mean “real opportunity”

Adecco’s analysis is explicit that the outcome of this technological shift is not predetermined. Several priorities stand out for leaders who want AI to produce inclusive, sustainable gains rather than patchy benefits.

1. Communicate a clear AI roadmap. Adecco recommends that organizations spell out how AI supports business priorities and creates opportunities for employees, rather than talking only about efficiency and cost savings. Articulating which tasks will change, which new roles will emerge, and how employees can move into them helps convert abstract fear into concrete, navigable change.
2. Engage workers early in redesigning roles. The study emphasizes involving employees in conversations about how their roles, skills, and career paths will evolve, moving away from top-down automation quietly implemented in the background. Co-designing workflows with workers and frontline managers can surface practical constraints and ethical concerns that pure technical planning might miss.
3. Invest in structured reskilling and transition pathways. Adecco and other sources consistently highlight targeted skills investment as a central lever for turning AI adoption into broad-based productivity and wage gains. That includes internal learning programs, partnerships with education providers, and clear transition routes from declining roles into growth areas such as AI operations, data analysis, and digital collaboration.
4. Strengthen governance, data transparency, and trust. Transparent AI governance—covering data use, bias mitigation, model oversight, and accountability—is critical to sustaining workforce trust as AI agents embed deeper into workflows. Using workforce data responsibly and communicating how decisions are made helps avoid the perception that AI is a black box used primarily to cut costs.

If these ingredients are missing, the default trajectory could still be one of fragmented benefits and concentrated harms, even in the absence of an outright jobs collapse.

Key takeaways and what to watch next

Several clear conclusions emerge from Adecco’s study and the wider 2026 evidence:

1. AI is transforming work more than it is eliminating it. The dominant pattern today is task reallocation and role redesign, with relatively modest net job losses and significant new roles in AI-related functions.
2. There is no empirical sign of an imminent employment collapse. Both Adecco and independent labor-market analyses show softening in some areas but nothing approaching systemic, AI-driven mass unemployment.
3. The real risk lies in uneven adaptation. Entry-level office roles, routine knowledge work, and some managerial layers face concentrated disruption, while care, service, skilled trades, and complex physical-world jobs remain relatively insulated for now.
4. Reskilling, governance, and leadership choices will determine whether AI’s impact is inclusive. Organizations willing to invest in skills, communicate clearly, and design hybrid human–AI workflows are already pulling ahead, while those treating AI as a narrow automation tool risk short-term savings at the cost of long-term capability and trust.
5. The next few years will be a stress test for institutions. As AI moves from pilots to deep integration, labor regulations, education systems, and corporate governance frameworks will be tested on their ability to handle churn without leaving large segments of workers behind.

In practice, Adecco’s message is not “don’t worry about jobs”; it is “worry about transitions.” AI is unlikely to erase employment, but it will keep redrawing the boundaries between tasks, roles, and skills. Whether that evolution becomes broadly inclusive, productive, and sustainable will depend less on the algorithms themselves and more on the decisions that business leaders, policymakers, and educators make in the years ahead.