Work and Skills in the Age of Artificial Intelligence.

On January 8, 2025, the fifth edition of the Future of Jobs report was released. Published every two years, its purpose is to track major technological, economic, and social trends in order to anticipate the global evolution of the labor market. The 2025 edition thus offers a comprehensive view of the economic, technological, geopolitical, and demographic trends that shape and will shape the global labor market over the 2025–2030 horizon; the occupations that are growing and those in decline; the most in-demand current and future skills; and finally, the strategies implemented by companies to cope with ongoing shifts—whether in employee training, work automation, talent recruitment, or internal mobility.

This publication is part of the Davos Forum—World Economic Forum (WEF)—a non-profit foundation created in 1971 by the German engineer and economist Klaus Schwab. It has become famous, among other things, for its annual meeting held each January in the Swiss town of Davos. The 2025 edition brought together more than 3,000 leaders from over 130 countries, including business executives, political figures, economists, intellectuals, scientists, academics, as well as representatives of major international organizations. The main topics addressed revolved around artificial intelligence (AI), geopolitical tensions, the ecological transition, and technological innovations.

The Future of Jobs 2025 is therefore a forward-looking document whose primary aim is to provide visibility to public decision-makers, companies, but also civil society and education stakeholders, so as to anticipate labor-market transformations in the age of artificial intelligence, climate transition, and geopolitical tensions. To that end, it draws on a database stemming from a large-scale survey conducted with more than 1,000 companies representing over 14 million workers, across 22 sectors and 55 national economies.

The document is structured in a logical way around five major parts whose purpose is to explain: (1) from a macroeconomic and systemic standpoint, the major forces currently at work that are transforming employment dynamics worldwide—technological, ecological, economic, geopolitical, and demographic; then (2), on the basis of these underlying “megatrends,” to study and understand the overall trajectory of employment as well as sectoral transformations; (3) to draw lessons from this with regard to the evolution of indispensable skills in a rapidly changing labor market; (4) to present the strategies and responses planned by companies in order to adapt their organization to the upheavals highlighted in the previous three parts; and finally (5), to offer a differentiated and more nuanced analysis of the global trends observed according to the major regions of the world, levels of economic development, and sectors of activity. This last part thus shows that ongoing transformations of the world of work—far from being uniform and simultaneous—also depend on structural factors specific to each context, such as national and regional economic conditions, the industrial structures of the territories studied, sometimes divergent demographic dynamics, but also political and ideological factors, such as public policies or social models.

In this article, I will focus solely on the technological macrotrend, considered as the principal driver of work transformation. I will examine its supposed effects on the reshaping of the employment market, the reconfiguration of tasks and the redefinition of the boundary between humans and machines that it implies, as well as the disruptions it generates for the evolution of skills through 2030.

A careful reading of the report raises a fundamentally important question, well known to philosophers of technology: that of the disruptions caused by the introduction of a new technical system—in this case, artificial intelligence—into the existing social system, including production structures, organizations, the education system, forms of knowledge transmission, political institutions, legal norms, cultural values, and ways of life. Yet our modernity, as the late Bernard Stiegler (1952–2020) never ceased to remind us, is characterized by a world in which the technical system is in perpetual transformation. In other words, the instability proper to the technical system—whose pace of transformation keeps accelerating—correlatively generates a structural instability of the social, which must constantly adjust to its evolution. This dynamic, which Clayton Christensen helped popularize in the 1990s under the term “disruption,” reveals a characteristic tension of our time: the phenomenon of permanent adjustment of human beings and social structures to the mutations and instability of the technical system. What the report calls “recomposition,” “resilience,” “flexibility,” “adaptation,” “lifelong learning,” etc., we should call more simply: “adjustment.”

AI, the leading driver of work transformation. – Among the macrotrends at work, the first factor of transformation for companies and the labor market highlighted by employers lies, in 60% of cases, in the widening of access to digital technology. This underlying movement is observed in a similar way across all regions of the world, whatever their level of development. In the list of nine key technologies selected by the Future of Jobs survey, three stand out as having the strongest expected impact: 1) artificial intelligence and information-processing technologies (86%); 2) robots and autonomous systems (58%); and finally, 3) energy generation, storage, and distribution technologies (41%). Artificial intelligence and information-processing technologies thus clearly come out on top among the technologies expected to have the greatest impact on employment and business activity by 2030.

The authors of the report rightly emphasize the predominant role played by generative AI since the launch of version 3 of ChatGPT in November 2022, which was accompanied by AI investment flows multiplied by nearly eight, notably to finance the deployment of infrastructures (data centers, servers, GPUs, compute clusters, storage, networks, energy production, etc.) necessary for training and deploying models. Despite this enthusiasm, AI—especially generative AI—still remains, to date, more a promise of productivity gains than a set of applications widely deployed and used by companies. Although rapid, its adoption has progressed unevenly depending on sectors and the level of economic development.

Over the longer term, moreover, several studies and observers mention a number of positive impacts on employees depending on their degree of specialization. While less specialized employees could, thanks to generative AI, carry out a greater variety of so-called “expert” tasks, more highly qualified professionals (electricians, doctors, engineers) could, for their part, access a more advanced level of knowledge, enabling them to solve more complex problems. This points more toward an augmentation of human capacities than toward their substitution. On the condition, however, the authors specify, that companies and public decision-makers establish incentive structures and public policies that steer technological development toward the enrichment of skills—otherwise there is a risk of seeing phenomena of replacement of human labor, associated with rising inequalities and unemployment.

Hence a strong growth in demand for training worldwide, as shown by enrollments in generative AI courses on the Coursera platform between 2022 and 2024, with a particularly marked acceleration after 2023, i.e., a few months after the launch of ChatGPT 3. Depending on whether they are driven by individual users or by companies, these courses cover topics that are more or less operational, as diverse as prompt engineering, trustworthy AI practices, AI-related strategic decision-making, the use of AI tools integrated into software such as Microsoft Excel, or the use of technologies that enable the development of applications.

A labor market in recomposition. - By cross-referencing information collected from survey respondents with real-world global employment data provided by the International Labour Organization (ILO), the report’s authors estimate that 22% of existing jobs at the time of writing should be affected by the macrotrends described in the first part of the report by 2030. More precisely, 170 million jobs should be created—i.e., 14% of global employment at end-2024—while 92 million should be eliminated, which would represent a net creation of 78 million jobs (7% growth). In other words, the authors estimate that one fifth (20%, i.e., one worker out of five) of current formal jobs—i.e., declared jobs governed by an employment contract—should be transformed, renewed, or reallocated in only five years. If these figures announce a significant phase of mutation of the labor market and employment, we are nevertheless very far from the catastrophist discourses predicting mass unemployment triggered by the rise of artificial intelligence and the automation process associated with its generalized deployment.

In other words, the main challenge posed by the transitions underway lies less in the destruction of jobs feared by proponents of the “end of work” than in the recomposition of the labor market, and in the issues that follow for public policies and companies, notably in terms of training, sectoral reconversion, and support for professional mobility. Yet, according to the executives surveyed, the fastest-growing roles by 2030 are precisely those driven by the development of new technologies and the expansion of digital access, particularly in the fields of AI and robotics. Among these roles, the report cites: Big Data specialists; FinTech engineers; artificial intelligence and machine learning specialists; software application developers; security management specialists; and data warehouse management specialists.

In addition to being the principal factors of transformation and net job creation by 2030, the expansion of digital technology, artificial intelligence and information-processing technologies—including robots and autonomous systems—also constitute the principal drivers of the destruction of routine, highly standardized jobs, and therefore jobs readily automatable by AI and robotization. Among the declining professions, the report notably mentions: accountants and auditors; hotel concierges and receptionists; telemarketers; legal secretaries and legal services staff; traveling salespeople; cashiers and ticket clerks; data entry operators; bank tellers and similar banking service employees; postal service workers, etc.

Digital technologies should thus represent the most divergent factor of change on the labor market, since they are expected both to create and to destroy more jobs than any other macrotrend mentioned in the introduction (19 million and 9 million respectively). Within this framework, the three underlying trends—1) the expansion of access to digital technology, 2) advances in AI, and 3) robots and autonomous systems—constitute, in reality, the three main growth drivers for the ten fastest-growing jobs.

The question of work reconfiguration. - The future of work therefore does not involve the disappearance of work, but rather an imperative of skills adaptation, increasingly oriented toward the use of digital technologies and, in particular, applications linked to artificial intelligence. In this context, it becomes urgent for companies: 1) to acculturate their employees to the use of AI; 2) to anticipate potential skill shortages; 3) to invest in internal and continuous training; 4) to consider from now on the organizational challenges associated with what one might call a fourth—more radical—phase of the digital transformation of organizations, after centralized computing of the 1960s–1980s (mainly used in large organizations such as banks, insurers, or public administrations); microcomputing from the late 1970s and especially the early 1980s, with the first graphical interfaces and the democratization of the personal computer (PC); the emergence of the Web in the early 1990s and its flagship technologies (browsers, search engines, e-commerce, etc.); the development of Web 2.0, social networks, mobile, and cloud architectures in the 2000s; then the rise of artificial intelligence from the 2010s, with machine learning, deep learning, frameworks such as TensorFlow or PyTorch, platforms such as Microsoft Azure, Google Cloud or Amazon Web Services, Transformer-type architectures, and finally generative artificial intelligence, which seems recently to be propelling organizations into a new stage of digital transformation, ultimately initiated as early as the beginning of the 1960s.

Hence perhaps the most important question, at the center of all debates, when it comes to addressing the real impacts of AI on employment and work: what about the boundary between humans and machines? Between what can be automated by artificial intelligence technologies, and what belongs to a know-how or a way-of-being that is typically human and not reproducible by a machine? More than any previous technological revolution, AI invites companies to redefine, in a quasi-systematic way, all job positions, or even to modify our ways of working, whether individually or collectively. It is an inexorable process, constantly questioned anew as AI technologies become increasingly versatile, and whose generalist character should only grow over time. In this framework, an ever larger share of the tasks we perform daily is destined to be automated by machines, or at least to be carried out in collaboration with them.

One can thus identify three types of tasks: 1) those performed by humans; 2) those mainly carried out by machines (algorithms); 3) those carried out by humans in collaboration with machines. It is this mix that should, according to survey respondents, evolve over the next five years. Indeed, while 47% of tasks are today carried out by humans alone, versus 22% by machines, and 30% by a combination of the two, employers anticipate an even distribution among the three by 2030. That is, a shift from 47% to 33% for exclusively human tasks; from 22% to 33% for tasks performed only by technology; and finally from 30% to 33% for human-machine collaboration.

In just under five years, we would thus witness, on a global scale, an average 50% increase in tasks performed exclusively by machines (from 22% to 33%), with a correlated decrease of around 30% in tasks performed by humans alone. In other words, over the 2025–2030 period, the human contribution to work would decrease by about 15 percentage points, of which 82% of this decline would be attributable to advances in automation, while only 19% would stem from an increase in human-machine collaboration. The report thus seems to confirm that the global economy has entered a phase of growing automation of work and production, due to the ever greater assumption by AI and robotics of repetitive and standardized tasks. Although general and sparing no industry, this process of automation or substitution of machine-work for human-work is not structurally homogeneous. If, for certain sectors such as insurance, pension management and telecommunications, more than 95% of the decline in human work comes from the automation of an ever larger number of tasks, in other sectors, such as medical services and healthcare, as well as public and government services, more than half of this reduction is due to a rise in human-machine collaboration.

Which seems to bring us back to our starting question: does the automation made possible by advances in artificial intelligence constitute a factor of elimination or of reconfiguration of human work? And, in the case of a reconfiguration—which by definition implies a significant decrease in the share of tasks performed exclusively by humans—what kind of transformation would work undergo? In other words, what other tasks would human work inherit? What is difficult, in fact, to assess today—and what the report pinpoints perfectly—is the effect that an increase in the share of tasks automated by machines (algorithms, AI applications, etc.) could have on the potential evolution of the total volume of tasks we might eventually perform, given the current configuration of work. The growing integration of artificial intelligence technologies into production processes would in reality generate two major phenomena: on the one hand, what is commonly called substitution effects, over a perimeter of existing tasks sufficiently repetitive and routine to be handled by machines, broadly understood; on the other hand, concomitant complementarity effects, with the emergence of new, so-called “more complex” tasks—or at least tasks difficult to automate—so that the total volume of work could increase.

Thus arises the double essential question raised by the growing use of artificial intelligence within organizations and the effects on the reconfiguration of human work that this use induces: both the quantity of new tasks likely to be created at the scale of society as a whole, and their value and cognitive interest. In other words: what are the expected effects of advances in AI on the volume and nature of jobs? And, in a context where an increasing share of the income generated by companies would come from AI systems and ever more advanced algorithms—something that the surveyed employers’ expectations seem to confirm—how do we ensure that a lasting share of the economic value created benefits human workers first and foremost? It is in this framework that the third category of tasks mentioned above, namely human-machine collaboration, takes on its full importance. Provided that we make, starting today, the choices that impose themselves, whether in terms of public policies, investments, or skill development strategies. The Upheaval of Skills over the 2025–2030 Period. – There is indeed no doubt that the technological advances of recent years, combined with the recent emergence of generative AI in November 2022—with the launch of ChatGPT—are forcing companies to question in depth the reconfiguration of tasks, operations, processes, and organizations that the integration and use of these tools imply, with a view to optimizing productivity gains, competitiveness, and employee satisfaction at work.

This reflection is coupled with another question, correlated with the first: that of the essential skills required in a work environment that should be increasingly augmented by applications and intelligent systems based on machine learning and generative AI. Organizations therefore face a double challenge: uncertainty as to the real impacts that AI in general, and generative AI in particular, could have in the medium and long term; and the level of disruption that these tools could cause to their employees’ skills.

It is therefore not surprising to note that, although slightly down, the pace of skills disruption expected by employers remains high. Overall, employers estimate that 39% of workers’ core skills should change by 2030. This figure was 44% in 2023, and 57% in 2020 (the Covid period). This favorable evolution, which reflects better visibility regarding the changes ahead and their anticipation, would be due to the efforts made by companies in terms of lifelong learning over the 2023–2025 period.

In addition to having durably entered a phase of growing automation, work would thus undergo a significant phase of mutation, translating into a structural updating of skills, driven notably by technological disruption and artificial intelligence. The figures mentioned, however, indicate only averages that conceal major disparities in the anticipated impacts that technological upheavals could have on skills. While countries such as Denmark (28%), the Netherlands (30%), the United Kingdom (33%), France (33%), China (33%), Germany (34%), or the United States (35%) anticipate disruptions below the average, economically less advanced countries such as Egypt (48%), Colombia (44%), or even Portugal (44%), Turkey (44%) or Israel (43%), show higher levels of transformation.

Given the framework we have just outlined, what are the essential (“core”) skills most valued by survey respondents in 2025? But above all, how does this skills framework evolve between 2025 and 2030? In other words, what is the trajectory of skills evolution between 2025 and 2030, and which skills will be most demanded by employers by 2030? One may moreover regret the absence of a definition, at least a generic one, of the skills mentioned in the report, even if a simple semantic understanding of the terms used may seem sufficient to form an idea of them—if not precise, at least spontaneous and consistent with common usage. One may also consider that a given skill—and even more so when it belongs to so-called “behavioral” skills—can take on different meanings depending on sectoral context but also, and above all, depending on cultural context.

Three skills acclaimed in 2025. - In 2025, the three core skills most acclaimed by employers are, in order: analytical thinking (69% of respondents); resilience, flexibility and agility (67%); and leadership and social influence (61%). These elements underscore both the importance of the capacity to understand, interpret, break down, prioritize, organize and connect information in order to solve problems logically and effectively—what one may understand, as a first approximation, by “analytical thinking”—and the ability to face difficulties, adapt to change, and inspire others, dimensions generally classified as behavioral and relational skills. Next come creative thinking (57%) and motivation combined with self-awareness (52%). Finally, one can mention a set of skills close in frequency: digital literacy (51%); empathy and active listening (50%); lifelong learning (50%); talent management (47%); and customer orientation (47%).

This set of ten skills reflects, in 2025, a clear trend on the part of employers to seek “hybrid” profiles combining technical skills (analytical thinking, creative thinking, mastery of digital tools), behavioral and relational skills (leadership, empathy, active listening), and skills related to personal development (self-awareness, motivation, curiosity and the willingness to learn throughout one’s life).

The trajectory toward 2030. - Yet an analysis of skills evolution as anticipated by employers by 2030 reveals significant changes, whose inflection owes much to ongoing technological and organizational transformations. Three major trends emerge. First, the growing importance of technological skills—AI and big data, networks and cybersecurity, digital literacy—which show the fastest growth across almost all sectors, with the exception of agriculture, forestry and fishing, as well as hospitality, catering and leisure. This underlying trend confirms the digital transformation of the global economy at work since at least the 1990s, but above all—and this is perhaps what is new—the anticipation of the automation of an ever larger number of tasks, whatever the nature of those tasks (cognitive or manual), with on the one hand industries where technology progresses rapidly, and on the other hand more traditional sectors that prioritize physical experience, direct interaction, and relational and behavioral know-how rather than technical skills. This sectoral polarization, which is also anthropological, fundamentally questions the nature of work and the still unclear criteria that will determine the redistribution of tasks and skills between humans and machines.

Second, the confirmation of cognitive and behavioral skills such as creative thinking, resilience, flexibility and agility, curiosity and lifelong learning, leadership and social influence, talent management, as well as empathy and active listening, which continue to grow in importance. These skills seem all the more valued and sought after by employers, 1) because they highlight qualities and aptitudes not yet fully reproducible by machines or artificial intelligence systems. In that sense, they refer to what is still, at this stage of AI development, typically human in work; 2) because the perception of their value is less intrinsic than relational, in the sense that they complement tasks now automated or automatable. In other words, by freeing up time, AI not only makes it possible to carry out high value-added tasks, but also helps strengthen cognitive and behavioral skills by acting as a personal assistant, consultant or coach; 3) because they point toward forms of know-how and ways of being that valorize adaptation to change and radical technological breaks, in a world that is increasingly complex, uncertain, and unpredictable.

Finally, the report mentions the strong growth of skills such as responsible environmental stewardship and systems thinking, which reflect both the growing importance of sustainable development and the need to conceive productive activities within a global framework integrating the complexity of interactions between the different elements of a system. In other words, it is a matter of going beyond analytical thinking alone—often limited to breaking a problem into distinct elements and seeking linear cause-and-effect relationships—in order to foster a more integrated and holistic approach, capable of embracing complexity and coping with the dynamic forces shaping the evolution of a system, whether economic, technological, societal, geopolitical or demographic.

Another remarkable fact highlighted by the report is the marked and anticipated decline in demand for skills related to reading, writing, mathematics, manual dexterity, endurance and precision, but also reliability and attention to detail. It is almost as if the recent development of generative AI—which took off at the time of the publication of the 2023 edition of the Future of Jobs report, which was supposed to account, among other things, for the impacts of technological advances on employment and skills over the 2023–2027 period—suddenly made the development of abilities as fundamental as mastery of language and calculation less important. The automation movement, which is not new since the first industrial revolution, would thus affect not only human manual and physical capacities, but also and above all cognitive capacities such as language knowledge, spelling, lexical, syntactic and grammatical skills, as well as logical, deductive and computational skills, and even the ability to reason and connect propositions.

These skills do not become obsolete or useless. However, the perception of their intrinsic value tends to diminish, insofar as it is now possible to delegate part of these tasks to generative AI systems whose linguistic performance—provided they are used appropriately and with due awareness—already surpasses that of most humans. This observation is corroborated by a study conducted by Indeed as part of the Future of Jobs Report 2025, which evaluated more than 2,800 skills according to their current level of substitutability by generative AI. On a five-level scale—very low, low, moderate, high, very high—69% of the skills analyzed showed low or very low substitutability, notably those considered “protected” because they are deemed deeply human, such as empathy and active listening. By contrast, the study highlights a markedly higher substitution potential for skills mobilizing theoretical knowledge and digital processing, such as data extraction or the use of machine learning models. The same holds for skills in reading, writing, mathematics and multilingualism, in tasks such as summarizing complex content, drafting texts, performing calculations, or translating. In other words, generative AI seems to make it possible, on the one hand, to automate a set of repetitive intellectual tasks and, on the other hand, to reinforce the relative value of typically human skills.

Four groups of skills. - The report thus distinguishes four groups of skills according to two dimensions: their current level of importance and their future trajectory toward 2030. In other words, while all the skills we are going to mention remain valued to some extent, this valuation varies in degree and timing given the ongoing transformations of the world of work. The proposed classification can thus be read as a strategic skills matrix that can serve as a basis for reflection for companies in implementing workforce and career pathway management (Gestion des Emplois et des Parcours Professionnels, GEPP)—which has replaced Gestion Prévisionnelle des Emplois et des Compétences (GPEC)—by integrating sectoral specificities and the organizational transformations associated with digital and ecological transitions. It also offers, for States and education systems as a whole, valuable indications for public education policy and the orientation of pedagogical supply.

• Core skills in 2030: A first group called “Key skills in 2030,” or “Core skills in 2030,” considered essential today and whose importance should continue to grow by 2030. Among these core skills, we find our mix of technical and theoretical skills centered on technologies and the development of individuals’ cognitive capacities, such as AI and big data, analytical thinking, creative thinking, systems thinking and digital literacy; and, on the other hand, behavioral skills centered on human qualities when individuals evolve in contexts of rapid technological progress, such as curiosity, lifelong learning, resilience, flexibility and agility, as well as leadership and social influence.

• Core skills in 2030 : Analytical thinking. Creative thinking. Systems thinking. Artificial intelligence and big data. Digital literacy. Curiosity and continuous learning. Resilience, flexibility and agility. Leadership and social influence. Talent management. Motivation and self-awareness.

• Emerging skills : A second group described as “emerging skills,” judged less essential today but whose importance, according to employers, is expected to grow by 2030. In other words, while they are not currently considered essential to the same extent as those in the first group, they are expected to play a critical role over the next five years. Among these skills are: networks and cybersecurity management, and environmental responsibility.

• Emerging skills : Networks and cybersecurity management. Environmental responsibility (ethics). Design and user experience (UX).

• Steady skills : A third group of so-called “stable” skills (“steady skills”), judged essential today but whose importance should not increase over the next five years. This category includes skills such as empathy and active listening, resource and skills management, as well as customer orientation and customer service. Although important and necessary, they are no longer sufficient to remain competitive in a world largely dominated by the mastery of technological know-how and the development of cognitive skills (analytical, creative and systems thinking).

• Steady skills : Empathy and active listening. Resource and skills management. Customer orientation and customer service.

Out of focus skills : Finally, a fourth group of skills judged less essential today and for which respondents do not anticipate any increase in importance by 2030. They include a whole network of so-called “technical” skills (“hard skills”), reflecting what is commonly called traditional areas of expertise, such as programming, marketing and media, training and support (coaching), mastery of foreign languages, as well as skills in comprehension, writing and mathematics.

·

• Out of focus skills : Programming. Marketing and media. Training and support (coaching). Mastery of foreign languages. Reading, writing and mathematics. Manual dexterity, endurance, precision. Reliability, attention to detail.

In a context of deep skills disruption mainly driven by the technological revolution linked to artificial intelligence, big data and new technologies, it is therefore not surprising to see that most industries—with the exception of a few such as agriculture and real estate—have begun to intensify their training effort, moving from 41% of employees having received training in 2023 to 50% in 2025, an increase of 22%. By 2030, this effort should be maintained, or even accelerated, since employers anticipate that around 60% of their employees will need to be trained, whether to remain competitive in existing positions or to be reassigned to new roles. This training investment by companies is mainly motivated by three key indicators: improved productivity (77% of respondents); increased competitiveness (70%); and finally talent retention (65%). This tends to show that lifelong learning and the continuous updating of skills are more than ever a sine qua non condition for remaining employed in the years to come.

Strategies for transforming skills and organizations. – This continuous updating of skills, in a context of growing instability of technological mutations and progress driven notably by advances in artificial intelligence—an instability perfectly illustrated since the launch of ChatGPT in November 2022, which has not only experienced an unprecedented adoption speed in the history of technologies, reaching 100 million users in only two months and then 800 million weekly active users in October 2025 (OpenAI source), i.e., nearly 10% of the world population, but whose capabilities have continued to grow since that date—comes together with the need, for companies, to rethink and deeply transform the organization of work. Not to mention that the continuous-improvement dynamics of artificial intelligence models has the effect of shifting the boundary between “human work” and “machine work,” constantly changing the perimeter of what it is possible—and economically preferable—to automate. This transformation thus leads to re-questioning what determines the local and global effectiveness and performance of an organization; the way jobs can be redefined; and the way specialized tasks can be coordinated and redistributed in a context in which humans will increasingly be called upon to play a supervisory role vis-à-vis technology (human-machine relations).

All these points refer to both the most strategic and the most operational aspects of organizational management, including the choice of means and actions to be implemented in order to achieve objectives, the determination of skill needs and the analysis of gaps relative to current situations, as well as the definition of priority human-resources strategies to be put in place by 2030. This in turn implies choices regarding internal policies for recruitment, skills assessment, talent management, training, development, mobility and professional transition, support and change management, compensation, well-being at work, etc. These are the central issues addressed in the fourth part of the report, which engage reflection at two levels: on the one hand at the local level, that of organizations; on the other hand at the national level, that of public policies regarding education reform, lifelong learning, employment policy, and labor policy.

In this framework, the report addresses a series of key themes, ranging from the identification of transformation obstacles as defined by survey respondents, to the definition of the strategies favored by companies to apprehend ongoing technological mutations and prepare individuals and organizations to work in environments deeply reconfigured by artificial intelligence. These are the various points that we will now review, in a non-exhaustive manner, retaining only those we consider the most essential.

Obstacles to transformation. – In response to current macrotrends, survey respondents identify two major obstacles to transformation for the 2025–2030 period: 1) a significant skills deficit in the labor market (63% of respondents); 2) organizational culture and resistance to change (46%). Among the other obstacles mentioned are: a regulatory framework judged obsolete or inadequate (39%); difficulties attracting talent, whether due to the attractiveness of the sector (37%) or of the organization itself (27%); a lack of adequate data and technical infrastructures (32%); insufficient investment capital (26%); as well as a still insufficient understanding of future opportunities. These results reveal—independently of regional variations—that current transformation challenges are less technical than they are above all human, cultural, organizational, and managerial. They clearly highlight the gap between the pace of evolution of artificial intelligence technologies and the capacity of organizations to evolve their modes of operation accordingly, in a context where they still struggle to perceive clearly the gains and opportunities they could derive from them.

Hence, compared to 2023 in particular, a rise in concern about the availability, recruitment, development and retention of talent—only 29% of companies expect an improvement in talent availability over the 2025–2030 period—which places organizations in front of a dilemma: on the one hand, the necessity to transform; on the other, the impossibility of doing so because of the shortage of talents and skills indispensable to driving change. Yet in a context where artificial intelligence tends to become a commoditized, standard, low-cost technology, companies’ competitiveness will rest more on their ability to integrate it quickly and effectively into their organizational processes, in order to fully reap the expected productivity gains.

Priority HR strategies. – In this context, what strategies do employers plan to adopt in order to respond to labor-market transformations induced by ongoing technological changes over the 2025–2030 period? On the basis of the information collected from the 1,000 surveyed companies across 22 sectors and 55 national economies, four priority directions emerge. The first consists in promoting upskilling among employees already in place, according to 85% of respondents. The second aims to increase the automation of processes and tasks, a strategy cited by 73% of respondents. The third concerns hiring new talent capable of bringing emerging skills deemed necessary by 70% of respondents. Finally, a fourth direction consists in strengthening the skills of the existing workforce through the use of new technologies, in a logic of augmentation and complementarity rather than substitution (63% of respondents).

To these three main levers of transformation are added other existing human-resources practices, such as reallocating employees in declining roles toward growing functions, according to 51% of respondents. More radical measures are also mentioned, such as reducing headcount for employees whose skills are no longer relevant relative to future needs (41% of respondents).

While these responses are naturally not mutually exclusive and testify to major organizational shifts to come, it is interesting to note that three main axes of transformation emerge. The first lies in a sustained effort of continuous training among existing employees, in order to enable the acquisition of new skills deemed essential by 2030. The second consists in reconfiguring existing tasks and processes with a view to amplifying the movement of work automation. The third, finally, aims to support employees in adopting and using emerging technologies so as to strengthen complementarity and synergy between humans and artificial intelligence systems.

All these orientations thus highlight a significant risk of labor-market polarization, with, on the one hand, employees possessing the skills necessary to integrate the use of new AI technologies—in other words, they would be the beneficiaries of the artificial intelligence revolution—and, on the other hand, a segment of employees whose skills would be automated or rendered obsolete, and who, consequently, could be marginalized or even excluded from the labor market if support or reskilling policies are not envisaged in time.

Increasing talent availability. – In a context where the main obstacles to transformation invoked by companies are less technological than human—scarcity of key skills, resistance to change, difficulty attracting talent, etc.—and where continuous training, people’s adaptability and the recruitment of new talent constitute priority strategies toward 2030, it is not surprising to note that improving talent availability appears as the main means of implementing these strategies. The central question thus becomes: what levers—HR, organizational, and public—are capable of ensuring access to the skills needed to accompany ongoing transformations, in an environment where technologies—and in particular artificial intelligence—constantly redefine expectations and needs? In other words: how can we increase the number of qualified employees? How can we retain existing talent? How can we support their progression? How can we foster the emergence and development of new talent? How can we retain key employees? How can we attract the indispensable external skills? These questions structure companies’ current thinking in the face of technological uncertainty and labor-market mutations.

If the range of good practices envisaged by organizations is broad and heterogeneous, one constant nevertheless emerges: the need to place the employee experience at the heart of talent-management strategies. Attractiveness, retention and upskilling no longer depend solely on recruitment or training policies, but on a global and integrated approach aimed at offering a stimulating, inclusive, balanced work environment, conducive to development, recognition and employee well-being.

Hence the importance granted—whatever the sector—to supporting health and well-being as the number-one priority for 64% of respondents, in order to strengthen talent availability over the 2025–2030 period. This priority thus rises from the 9th place it held in the 2023 edition to the 1st place in the 2025 edition. Among the other means put forward to improve talent availability, one can cite, in chronological order: the implementation of reskilling and upskilling programs for employees (63%); the improvement of progression pathways and talent-promotion processes (62%); offering higher salaries (50%); tapping into underutilized talent pools (47%); opportunities for remote and hybrid work (43%); the implementation of diversity, equity and inclusion policies and programs (39%); improving policies on working hours and overtime (38%); or the clear articulation between the company’s purpose and its positive impact on society, the environment and its customers (37%).

This shift toward an economy and a labor market in which a certain form of scarcity of key human skills makes it imperative for companies to become socially, culturally and economically attractive reveals another phenomenon, which three McKinsey consultants— Ed Michaels, Helen Handfield-Jones and Beth Axelrod—popularized in a book published in 2001: The War for Talent. While the war for talent has always existed, and companies have always sought to attract the best possible profiles in order to be more competitive than their rivals, it would seem that it has taken on an entirely different dimension with the recent development of AI and the gigantic financial and economic gains that its multiple applications make it possible to envision in the medium and long term. In other words, what seems new today is that this war for talent—which consists in securing the best possible profiles across all occupations related to the development of artificial intelligence—data analyst, data manager, data miner, AI developer, machine learning engineer, AI architect, prompt engineer, chief AI officer, AI project manager, data scientist, AI consultant, AI ethicist, etc.—has taken on an unprecedented scale and intensity, to the point of affecting, beyond the Tech giants waging a war with billions of dollars, all companies, whatever their size or sector of activity.