Artificial intelligence could reshape nearly one in four occupations across the Arab world by 2035 while adding hundreds of billions of dollars to regional output. Whether that becomes a productivity dividend or a labour-market shock will depend on how quickly education, skills and social protection catch up with investment.
Artificial intelligence is advancing across the Arab world faster than many of its labour markets are preparing for it.
Nearly 25% of occupations in the region could be affected by generative AI, through either displacement or augmentation, while mismatches between education systems and labour-market requirements already reach 40%-70% in some Arab countries, according to a joint report by the United Nations Economic and Social Commission for Western Asia and the International Labour Organization.
The contrast points to the central economic challenge confronting the region.
Governments and companies can build data centres, buy computing capacity and deploy AI systems relatively quickly. Developing the workers, institutions and businesses capable of using that technology productively takes considerably longer.
The result is an emerging race between AI capital and workforce capability — one that could determine not only which jobs disappear or change, but where the economic gains ultimately accrue.
The ESCWA–ILO report, Artificial Intelligence and Employment Futures for the Arab Region, examines how AI could reshape employment between 2025 and 2035. Rather than making a single forecast, it sets out three scenarios showing how different combinations of technological adoption, skills investment and policy preparedness could lead to sharply different outcomes for growth, jobs and inequality.
Its broader message is that AI is likely to do more than eliminate jobs. It will reprice skills, reorganise tasks and redistribute productivity across workers, companies and economies.
Jobs Will Change Before Many Disappear
The first disruption is likely to be concentrated in routine work.
Clerical, administrative and repetitive occupations face some of the greatest automation pressure, while new opportunities are expected to emerge in education, healthcare and technology-intensive activities.
That makes the impact of AI more complicated than counting jobs created against jobs destroyed.
An accountant may remain an accountant while automating reconciliation and document analysis. A lawyer may use AI to search large volumes of material. A customer-service operation may employ fewer agents while enabling those who remain to handle substantially more inquiries. Doctors, teachers, engineers and financial analysts may become more productive without their professions disappearing.
The economic divide could therefore increasingly emerge within occupations: between workers who can use AI effectively and those performing similar functions without those capabilities.
AI does not have to eliminate a profession to change its economics. If one employee equipped with new technology can perform work that previously required several people, output per worker rises even when the occupation survives.
Productivity gains and labour displacement can therefore occur simultaneously.
AI Arrives on Top of a Skills Deficit
For Arab economies, that transition begins before an older structural problem has been resolved.
The report says mismatches between education systems and labour-market needs already reach 40%-70% in some countries, underscoring the scale of the existing skills gap.
AI risks widening it because technology can change job requirements much faster than universities can rewrite curricula, vocational systems can retrain workers or governments can redesign labour policy.
A degree programme may take years to reform. A new AI system can materially alter a job within months.
That changes the economic role of education.
The traditional model — acquire qualifications early in life, enter an occupation and rely on those skills for much of a career — becomes increasingly difficult to sustain when the technology embedded in work changes continuously.
ESCWA and the ILO therefore place strong emphasis on lifelong learning, ranging from basic AI literacy across the workforce to advanced technical, analytical and governance capabilities. The report also highlights distributional impacts across age and gender and risks of deeper exclusion for older workers, persons with disabilities, migrants and refugees.
The requirement extends far beyond producing more software engineers.
The larger productivity opportunity lies in enabling doctors, teachers, bankers, accountants, engineers, civil servants, logistics managers, manufacturers and entrepreneurs to use AI inside existing sectors.
Skills policy therefore becomes economic infrastructure.
Three Very Different Arab Economies in 2035
The report’s three scenarios demonstrate how widely outcomes could diverge.
They are not forecasts. They are alternative futures designed to test what happens when technological adoption, training and public policy move at different speeds.
Rapid AI, Rapid Workforce Adjustment
In the first scenario, rapid AI adoption is accompanied by strong investment in workforce transition and skills development.
Under this pathway, economies absorb substantial disruption while using the technology extensively across industries. The ILO says AI could contribute nearly 25% of regional GDP by 2035, despite millions of jobs being displaced and labour markets undergoing major restructuring.
The apparent contradiction is important.
Strong GDP growth does not imply limited employment disruption. The economic attraction of automation often lies precisely in allowing businesses to produce more with different combinations of labour and capital.
The policy question therefore becomes whether workers leaving declining tasks can move rapidly enough into new or expanding activities.
Technology Runs Ahead of Workers
The second scenario provides the report’s sharpest warning.
AI adoption accelerates, but education, retraining, labour-market institutions and social protection fail to keep pace.
Even then, AI could contribute an estimated $375 billion to regional GDP by 2035. At the same time, large numbers of lower- and middle-skilled workers could face prolonged displacement and downward occupational mobility, while inequality and labour-market fragmentation increase.
This is perhaps the report’s most consequential economic finding.
AI can raise output while producing a poor labour-market outcome.
GDP growth alone would therefore be an incomplete measure of whether the transition succeeded.
Governments would also need to examine employment participation, wages, productivity per worker, occupational mobility and whether the gains from new technology are distributed broadly or become concentrated among highly skilled labour and capital owners.
A More Measured Transition
The third scenario assumes more gradual AI integration combined with coordinated policy action.
That provides institutions and workers more time to adapt while still generating productivity gains and new opportunities, particularly in sectors including tourism, logistics and digital services.
Under this pathway, the contribution of AI to Arab economies could grow by 20%-34% annually, according to the report’s scenario analysis.
The underlying choice is therefore not simply between adopting AI quickly or slowly.
It is whether economies can build the complementary skills and institutions required to absorb it.
There Will Not Be One Arab AI Transition
The region enters this transformation from very different starting points.
Capital-rich economies can invest aggressively in data centres, cloud infrastructure and advanced technology partnerships. Populous middle-income states must simultaneously create employment at scale. Lower-income and conflict-affected economies face more fundamental constraints involving infrastructure, financing and institutional capacity.
Those differences could widen an existing inter-Arab productivity gap.
Economies able to combine capital, technology, skills, entrepreneurship and regulatory capacity may build higher-value AI-enabled industries.
Others risk becoming principally consumers of AI systems developed elsewhere.
That distinction matters because owning computing infrastructure or gaining access to advanced models does not by itself determine where economic value is ultimately retained.
The larger prize lies in building domestic companies and workforces able to use those technologies to increase output across finance, healthcare, manufacturing, logistics, education, tourism, public administration and professional services.
The Distribution Question
The transition will not affect all workers equally.
The ESCWA–ILO report identifies different effects across demographic groups and warns that technological change could deepen exclusion for workers already facing barriers to employment or training.
Highly skilled employees may be better placed to use AI as an augmentation tool, raising their productivity and potentially their earnings.
Workers concentrated in routine information-processing activities face greater substitution pressure.
The result could be labour-market polarisation: highly productive workers and firms capturing increasingly large gains while others move toward lower-value or less secure employment.
That makes job quality as important as job quantity.
An employee may technically remain in work yet experience weaker bargaining power, reduced hours, more monitoring or movement into less secure forms of employment.
Conversely, AI could expand employment access in some areas by enabling remote work, reducing physical barriers and supplementing language or technical capabilities.
Technology does not determine those outcomes by itself. Labour-market institutions help determine how its gains and costs are distributed.
Social Protection Becomes Technology Policy
The transition also challenges social-protection systems built around relatively stable employment.
AI could increase occupational switching, periods of retraining, platform work and movement between formal and informal employment.
Workers affected by technological change may therefore require support not merely while unemployed, but while acquiring new skills and moving between occupations.
The ILO calls for sustained investment in skills and lifelong learning, stronger labour-market institutions, expanded social protection and closer cooperation among governments, employers and workers’ organisations.
The economic principle is straightforward.
If companies capture productivity gains immediately while workers bear most of the cost of retraining and income disruption, adjustment is likely to be slower and the distribution of AI’s benefits less even.
Retraining and social protection therefore become part of the infrastructure of technological adoption rather than merely compensation for its consequences.
The Real Race Is Between Capital and Capability
The first phase of the Arab AI boom has largely been measured in capital: data centres, computing infrastructure, technology partnerships and investment commitments.
The next phase will increasingly be measured in productivity.
That changes the benchmark for success.
The decisive question is not how many AI projects governments announce or how much computing capacity economies install. It is whether businesses and workers use that capacity to create more sophisticated goods and services, raise output per worker and build competitive domestic enterprises.
Compute is capital. Productivity depends on whether workers and firms can use it.
That is where the timing problem becomes critical.
Computing capacity can be purchased. Models can be licensed. Data centres can be built.
Education reform, workforce retraining and institutional adjustment move more slowly.
Unless those processes begin to converge, Arab economies could reach 2035 with far greater technological capability but still lack enough of the human capability required to exploit it fully.
The ESCWA–ILO report therefore shifts the debate beyond whether AI will take Arab jobs.
The more consequential question is what happens to the productivity, incomes and career paths attached to the work that changes.
The region’s most successful AI economies may not necessarily be those that spend the most on infrastructure.
They will be those able to convert that investment into higher worker productivity, new high-value employment, competitive domestic companies and economic value retained at home.
That is the real test of the Arab AI transition.
Related news:
Global Debt Tops $365tn as AI Adds to Funding Demands
Egypt, Nvidia Move Towards AI Computing and Skills Partnership
Read also:
Algeria Starts Trans-Saharan Gas Link to European Export Network
Algeria Tenders Mobile Coverage for 3,000 Remote Communities
