AI and the Future of Work: Who Is at Risk, Who Benefits and How We Can Manage the Transition
Artificial intelligence is rapidly changing the world of work.
What once seemed like a distant technological possibility has become part of everyday business. AI can write reports, analyse documents, generate images and video, answer customer enquiries, translate languages, write computer code, process data and perform an expanding range of tasks that were previously carried out exclusively by people.
This has created enormous opportunities for businesses and workers. Companies can become more productive, customers can receive faster service, and entirely new industries and occupations are emerging.
But there is another side to the transformation.
If AI allows one employee to perform the work that previously required five, what happens to the other four?
This is the central economic and social question surrounding artificial intelligence. The issue is not simply whether AI will eliminate jobs. It is who will lose work, who will gain opportunities, how quickly the transition will happen, and whether the benefits of higher productivity will be shared fairly.
Current evidence suggests that the future is likely to involve a mixture of automation, job transformation and job creation rather than a sudden disappearance of human work. The International Labour Organization’s 2025 research found that roughly one in four workers globally are in occupations with some exposure to generative AI, but that transformation of jobs is more likely than outright replacement.
Nevertheless, the disruption could be substantial, particularly for workers whose jobs consist largely of repetitive, predictable digital tasks.
The challenge for governments, businesses and workers is therefore to ensure that AI becomes a source of shared prosperity rather than widespread insecurity.
Which Workers Are Most at Risk?
AI does not threaten every occupation equally.
The greatest risk is generally found where work consists of tasks that are repetitive, predictable, highly digitised and capable of being performed by software.
This is an important distinction.
A job is not necessarily going to disappear simply because AI can perform some of its tasks. In many cases, the job itself will change.
The ILO’s latest research identifies clerical occupations as having the highest exposure to generative AI. Among the occupations specifically highlighted are data-entry clerks, typists, accounting and bookkeeping clerks and administrative secretaries. Exposure has also increased in some professional and technical occupations, including financial analysts, programmers and web and multimedia developers.
1. Clerical and Administrative Workers
This is probably the clearest area of vulnerability.
AI can increasingly perform or assist with:
- Data entry
- Document preparation
- Scheduling
- Transcription
- Basic correspondence
- Record management
- Form processing
- Routine reporting
- Basic bookkeeping
- Administrative research
A company that once required several people to process documents may eventually need fewer employees because AI can handle much of the routine workload.
This does not necessarily mean that all administrative workers will be unemployed. Their roles may shift towards coordination, customer interaction, problem-solving and oversight.
But fewer people may be required to perform the routine component of the work.
2. Customer Service and Call-Centre Workers
AI-powered chatbots and voice systems are becoming increasingly capable of handling routine enquiries.
Customers can ask questions about:
- Opening hours
- Orders
- Account balances
- Delivery status
- Product specifications
- Refund policies
- Appointments
AI can answer many of these questions instantly and around the clock.
This creates a strong incentive for companies to reduce the number of people handling routine customer interactions.
Human employees are likely to remain important for complex, emotional or unusual cases, but the size of traditional call-centre operations could decline.
The ILO has specifically identified customer service and other clerical support roles as areas where AI-driven automation could produce employment losses.
3. Basic Accounting and Bookkeeping
Accounting is another area undergoing substantial change.
AI-powered software can increasingly:
- Categorise transactions
- Generate invoices
- Reconcile accounts
- Prepare reports
- Detect anomalies
- Extract information from receipts
- Assist with tax calculations
This is unlikely to eliminate the need for accountants altogether.
Businesses will still require professionals capable of interpreting financial information, advising management, dealing with regulators and handling complicated cases.
But the number of people needed to perform basic bookkeeping and data-processing tasks could decline.
4. Content Writers and Basic Copywriters
Generative AI can already produce reasonably competent:
- Blog articles
- Product descriptions
- Social media posts
- Advertising copy
- Email drafts
- Summaries
- Basic reports
This creates particular pressure on entry-level writing jobs.
The greatest risk is not necessarily to highly specialised writers, investigative journalists or experts with distinctive voices. It is to large volumes of relatively standardised content.
The same principle applies to translation, proofreading and routine editing.
People with deep subject knowledge, creativity, editorial judgement and the ability to produce genuinely original work are likely to remain valuable.
5. Graphic Designers and Production Artists
AI image-generation systems can produce advertising concepts, illustrations, product imagery and visual variations in seconds.
This creates pressure on work involving:
- Simple advertisements
- Basic social media graphics
- Background removal
- Image resizing
- Template-based designs
- Basic illustrations
- Routine production work
But design itself is unlikely to disappear.
Businesses still need people who understand branding, culture, typography, composition, strategy and the psychology of visual communication.
The designer’s role may therefore move from producing every individual asset manually to directing, refining and managing AI-assisted production.
This is an important distinction for creative agencies.
6. Programmers and Web Developers
Perhaps surprisingly, software development is also being transformed.
AI can generate code, explain code, identify bugs, create tests and assist developers with increasingly complex programming tasks.
The ILO’s 2025 research specifically notes increased exposure among application programmers and web and multimedia developers.
This does not mean that programmers are about to disappear.
Instead, the productivity of an experienced developer can increase significantly.
The potential consequence is that companies may require fewer developers for some projects while simultaneously expecting each developer to accomplish more.
Entry-level programming positions may face particular pressure because some tasks traditionally given to junior developers can increasingly be automated.
7. Marketing and Advertising
AI can analyse customer data, generate advertisements, segment audiences, create campaigns and optimise content.
This could reduce demand for some routine marketing functions.
But marketing also involves strategy, brand positioning, relationships, creativity and understanding human behaviour.
The marketer who knows how to use AI effectively may therefore replace the marketer who refuses to adapt.
8. Banking, Finance and Insurance
Financial services are particularly suitable for AI because they involve enormous quantities of structured information.
AI can assist with:
- Fraud detection
- Credit analysis
- Customer service
- Financial research
- Risk assessment
- Document processing
- Compliance monitoring
- Investment analysis
The ILO has identified financial analysts and investment advisers among professional occupations experiencing increased exposure to generative AI capabilities.
Again, the likely result is not that finance disappears, but that fewer people may be required for some routine analytical functions.
Who Is Less Exposed?
AI exposure is not evenly distributed across the economy.
Jobs involving substantial physical activity, unpredictable environments, human relationships or complex practical judgement are generally harder to automate with software alone.
These include many occupations in:
- Construction
- Agriculture
- Skilled trades
- Healthcare
- Personal care
- Hospitality
- Maintenance and repair
- Transport and logistics
- Childcare
- Social services
This does not mean these sectors are immune.
Robotics, autonomous vehicles, computer vision and other forms of automation can eventually affect many physical jobs.
But the path is generally more complicated than simply deploying software.
The ILO’s analysis shows that exposure to generative AI is substantially lower in low-income economies partly because a larger share of employment is concentrated in occupations such as agriculture, transport and food vending that are less exposed to this particular technology.
AI Will Also Create Jobs
The debate about AI often focuses almost exclusively on jobs that may disappear.
That is only half the story.
AI is also creating demand for new skills and occupations.
The World Economic Forum’s Future of Jobs Report 2025 estimates that, across the broader set of technological, economic and demographic trends affecting employment, approximately 170 million jobs could be created globally by 2030 while 92 million could be displaced, resulting in a net increase of about 78 million jobs. These figures are not an AI-only forecast, but they illustrate the scale of labour-market transformation expected from multiple forces.
Among the fastest-growing roles identified by the WEF are AI and machine-learning specialists, big-data specialists, fintech engineers, software and applications developers, data-warehousing specialists and security-management specialists.
Other opportunities will emerge in fields that are difficult to predict today.
Just as the internet created occupations that barely existed several decades ago, AI will create new forms of work.
The Biggest Risk May Be Unequal Opportunity
The most serious problem may not be that AI eliminates all jobs.
It may be that AI divides workers into winners and losers.
A highly skilled worker who knows how to use AI may become considerably more productive.
A worker whose role consists almost entirely of tasks AI can perform may find that their bargaining power declines.
The result could be greater inequality between:
- Highly skilled and low-skilled workers
- Large and small businesses
- Technology-rich and technology-poor countries
- Workers who receive training and those who do not
- People with access to digital infrastructure and those without it
The ILO estimates that 34% of employment in high-income countries is in occupations with some degree of GenAI exposure, compared with about 11% in low-income countries.
That lower exposure in poorer economies may initially appear beneficial.
But there is a danger.
If developing countries are slow to adopt AI, their workers may avoid some displacement while also missing the productivity gains that could help their economies become more competitive.
This creates a difficult balancing act.
The Gender Dimension
AI’s impact is also likely to be unequal between men and women.
Women are disproportionately represented in many clerical and administrative occupations, which are among the occupations most exposed to generative AI.
The ILO reported in 2026 that female-dominated occupations are almost twice as likely to be exposed to GenAI as male-dominated occupations—29% compared with 16%. It also found that women remain underrepresented in AI-related occupations.
This means that simply telling workers to “learn AI” is not enough.
Governments and businesses must ensure that women have equal access to the training, technology and career pathways leading to emerging AI-related occupations.
The Pros of AI in the Workplace
Despite the risks, the potential benefits are enormous.
Higher Productivity
AI can allow employees to complete routine tasks much faster.
A report that once took hours could potentially be produced in minutes.
This can make businesses more competitive and potentially reduce prices for consumers.
Better Services
AI can provide customers with assistance at any hour.
It can also reduce waiting times and help businesses respond more quickly.
New Businesses and Industries
AI is creating opportunities for entrepreneurs to build products and services that were previously impossible or prohibitively expensive.
Small companies can now access capabilities that once required large teams.
Reduced Human Error
AI can assist with repetitive calculations, document checking, data processing and pattern recognition.
Used properly, this can reduce certain types of mistakes.
More Meaningful Work
If AI takes over repetitive tasks, workers may have more time for activities requiring judgement, creativity, empathy and human interaction.
This is potentially one of the greatest benefits of automation.
The problem is ensuring that the productivity gains are actually used to improve work rather than simply eliminate workers.
The Cons and Risks
The disadvantages deserve equal attention.
Job displacement
Some workers genuinely will lose their jobs.
The transition will not be painless.
Wage pressure
Even when jobs remain, workers may find that fewer people are required to perform them, potentially weakening bargaining power.
Loss of entry-level opportunities
This could become particularly serious.
Many professions traditionally train new workers by giving them routine tasks.
If AI performs those tasks, young people may struggle to obtain the experience needed to progress into senior positions.
Greater inequality
Those who own AI systems and those who know how to use them effectively may capture a disproportionate share of the economic gains.
Surveillance and algorithmic management
AI can also be used to monitor workers, measure productivity and make decisions about hiring or performance.
Used irresponsibly, this could create highly stressful workplaces.
Loss of human skills
If people become excessively dependent on AI, some skills may gradually deteriorate.
Writing, calculation, research, problem-solving and decision-making still require human understanding.
What Can Be Done?
The answer should not be to stop AI.
Trying to prevent technological progress would probably be unrealistic and could deprive societies of enormous benefits.
The better objective is to manage the transition deliberately.
1. Make Reskilling a Normal Part of Employment
Businesses should treat employee training as an investment rather than a cost.
Instead of laying off workers whose tasks are becoming automated, companies should first examine whether those workers can move into other roles.
An administrative employee, for example, might be trained to manage AI systems, customer relationships, digital records or business operations.
A graphic designer might become an AI-assisted creative director.
A customer-service employee might become an AI-support specialist responsible for handling complex cases and supervising automated systems.
The question should be:
“How can this employee’s role evolve?”
before asking:
“How can this employee be replaced?”
2. Introduce AI Transition Plans
Large organisations adopting AI should be encouraged—or in some circumstances required—to assess the employment consequences before major automation projects are implemented.
A transition plan could include:
- Which jobs are affected?
- Which tasks are being automated?
- Which employees can be retrained?
- What new positions will be created?
- How long will retraining take?
- What happens to workers who cannot transition?
This creates greater transparency and gives employees time to prepare.
3. Share Productivity Gains
If AI enables a company to produce twice as much with the same workforce, the benefits should not automatically go entirely to shareholders.
Companies could use productivity gains to support:
- Higher wages
- Reduced working hours
- Employee training
- Better working conditions
- Profit-sharing
- New business expansion
A particularly interesting possibility is a gradual reduction in working hours.
If technology makes workers substantially more productive, society could choose to convert part of that productivity into more leisure rather than simply fewer workers.
4. Strengthen Social Protection
Workers who genuinely lose employment need a safety net.
This can include:
- Unemployment support
- Retraining grants
- Career counselling
- Job-matching services
- Portable benefits
- Support for entrepreneurship
The objective should not be to keep people permanently dependent on government assistance.
It should be to give them enough stability to move successfully from one form of work to another.
The ILO has repeatedly emphasised social dialogue and proactive policies for managing the transition fairly.
5. Reform Education
Education systems need to change.
Teaching people to memorise information that AI can retrieve instantly is becoming less valuable.
Education should increasingly emphasise:
- Critical thinking
- Communication
- Problem-solving
- Creativity
- Collaboration
- Digital literacy
- Entrepreneurship
- AI literacy
- Adaptability
Students should learn not merely how to use AI, but how to question its answers, verify its output and work alongside it.
6. Protect and Develop Human Skills
Some capabilities are likely to become more valuable precisely because AI cannot easily reproduce them.
These include:
Empathy.
Leadership.
Judgement.
Trust-building.
Negotiation.
Creativity.
Physical dexterity.
Human relationships.
Ethical decision-making.
Businesses should therefore avoid building workforces consisting entirely of technical skills.
The future may belong to people who combine technological competence with strong human capabilities.
7. Help Small Businesses Adopt AI
AI should not become a technology available only to large corporations.
Governments and industry organisations can help small and medium-sized businesses access affordable AI tools, training and digital infrastructure.
This is particularly important in developing economies.
If large companies adopt AI while millions of small businesses remain excluded, productivity and income inequality could widen.
Affordable access to AI could instead allow small businesses to compete more effectively.
8. Create Rules for Responsible AI
Governments have an important role in establishing basic rules.
Workers should know when AI is being used to make decisions that affect them.
Companies should be accountable for discriminatory or harmful automated decisions.
Personal data should be protected.
AI should not become a mechanism for removing human accountability.
Regulation should, however, avoid making AI so expensive or complicated to adopt that only the largest corporations can afford it.
The goal should be responsible innovation, not innovation paralysis.
What Should Businesses Do Now?
Businesses do not need to wait for governments to act.
They can begin by conducting an AI workforce audit.
For every department, identify:
- Which tasks are repetitive?
- Which tasks can AI assist with?
- Which tasks could eventually be automated?
- Which human capabilities remain essential?
- What new skills will employees need?
- Which existing employees could be retrained?
This approach is much better than simply asking:
“How many people can we replace?”
The better question is:
“How can we use AI to make our people and our organisation more productive?”
What Does This Mean for Developing Countries?
Developing countries face a particularly complicated challenge.
They have large young populations entering the labour market, significant informal employment and limited social protection in many sectors.
At the same time, they have an opportunity to use AI to overcome some of the disadvantages associated with limited resources.
AI could help small businesses with accounting, marketing, translation, customer service and research.
It could help teachers prepare educational materials.
It could help farmers analyse information.
It could help health workers with administrative tasks.
It could help governments process information more efficiently.
But infrastructure remains critical.
The ILO notes that inadequate electricity, broadband, digital skills, technology costs and operational difficulties can prevent countries from fully benefiting from AI.
The developing world therefore needs to pursue two objectives simultaneously:
protect workers from disruptive technological change while ensuring that they are not excluded from its benefits.
The Future: Human + AI Rather Than Human vs AI
The most useful way to think about the future may not be “AI versus humans.”
It may be:
“Humans using AI versus humans who do not.”
A lawyer using AI may outperform a lawyer who refuses to use it.
A designer using AI may produce more work than one who relies entirely on manual production.
A business owner using AI for customer service, marketing and analysis may compete more effectively with a larger company.
A teacher using AI intelligently may have more time to work individually with students.
This suggests that the future labour market may increasingly reward AI-assisted human capability.
The technology becomes a tool.
The human provides judgement, responsibility, creativity and context.
A Social Contract for the AI Age
Ultimately, the AI transition is not just a technology question.
It is an economic and social question.
Who owns the technology?
Who benefits from the productivity gains?
Who pays for retraining?
What happens to displaced workers?
How do we prevent inequality from widening?
How do we ensure young people can enter professions where AI is already performing traditional entry-level tasks?
And how do we ensure that developing countries are participants in the AI economy rather than merely consumers of technology created elsewhere?
These questions require governments, businesses, workers, educators and technology companies to work together.
There is no single policy that will solve the problem.
Conclusion: Do Not Stop the Future—Prepare People for It
Artificial intelligence will almost certainly eliminate some jobs.
It will also transform many more.
And it will create entirely new forms of employment.
The greatest mistake would be to assume that the choice is simply between embracing AI and protecting jobs.
We can do both.
Businesses can automate responsibly while retraining employees.
Governments can encourage innovation while strengthening social protection.
Schools can teach AI literacy alongside critical thinking.
Workers can continuously develop new skills.
And the gains from higher productivity can be shared through better wages, shorter working hours, new opportunities and investment in people.
The ILO’s latest evidence provides an important perspective: about one in four workers globally are in occupations with some exposure to generative AI, but the dominant expected outcome is job transformation rather than wholesale replacement.
That distinction matters.
The future of work does not have to be a future without work.
The real challenge is to ensure that when machines become capable of doing more, people are given the opportunity to do better work.
AI should not be viewed simply as a machine for cutting labour costs.
It can become a tool for increasing human productivity, creating new businesses, improving services and raising living standards.
But that outcome will not happen automatically.The technology will shape the future—but the choices we make about it will determine whose future it becomes.
Leave a Reply