Blog 03 - Will AI Replace Employees—or Redesign Their Jobs?

Will AI Replace Employees—or Redesign Their Jobs?

Few workplace questions create as much anxiety as the question, “Will Artificial Intelligence take my job?” The concern is understandable. Generative AI can now produce text, analyse information, generate code, support customer service, prepare documents and perform tasks that were previously associated with professional and administrative work.

However, the impact of AI on employment is more complex than a simple story of humans being replaced by machines. Jobs are made up of many different tasks. Some tasks can be automated, some can be supported by AI, some may disappear, and completely new tasks may emerge. This means that the more useful HRM question may not be “Which jobs will AI remove?” but rather:

How should organisations redesign jobs so that technology improves productivity without reducing meaningful work, employee capability and job quality?

Exposure Does Not Mean Replacement

The International Labour Organization’s updated global study of generative AI provides an important starting point. Gmyrek et al. (2025) estimate that approximately one in four workers worldwide are employed in occupations with some degree of exposure to generative AI. However, only 3.3% of global employment falls into the highest exposure category.

Figure 1. Potential exposure of global employment to generative AI. Author’s illustration based on Gmyrek et al. (2025), International Labour Organization.

The word “exposure” is important. It does not mean that every exposed employee will lose their job. It means that at least some of the tasks inside the occupation could potentially be affected by the technology. The ILO therefore argues that transformation of jobs is generally more likely than the complete disappearance of occupations.

This task-based perspective is useful for HR professionals because organisations do not employ abstract “occupations”; they employ people to perform combinations of activities. If AI can complete three tasks out of ten, management must decide what happens to the remaining seven tasks and how the employee’s role should change.

Automation and Augmentation Are Different

Two concepts are particularly important when discussing AI and work: automation and augmentation.

Automation occurs when technology performs a task that was previously carried out by a human worker.

Augmentation occurs when technology helps a human perform the task faster, better, more safely or with greater access to information.

The distinction matters because the same technology can be implemented in different ways. A company can introduce AI to remove employees from a process, or it can introduce AI to strengthen employees’ capabilities. The organisational outcome depends on management choices, work design, skills and strategy.

For example, an AI customer-service system could completely automate basic enquiries while human employees deal with more complex cases. Alternatively, it could operate beside employees, suggesting possible responses while the employee remains responsible for the conversation. Both models use AI, but they create very different jobs.

Real Evidence: AI Can Improve Human Performance

A large workplace study by Brynjolfsson, Li and Raymond (2025), published in The Quarterly Journal of Economics, examined the introduction of a generative AI assistant to 5,172 customer-support agents. The AI system did not replace the employees. Instead, it monitored customer conversations and suggested responses that employees could accept, modify or ignore.

Figure 2. Productivity effects of AI assistance in customer support. Author’s illustration based on Brynjolfsson, Li and Raymond (2025).

The researchers found an average 15% increase in issues resolved per hour. More strikingly, less-skilled and less-experienced employees achieved productivity improvements of around 30%. The study also found evidence that employees learned from the AI’s recommendations over time.

This finding challenges the assumption that AI always increases inequality between high-skilled and low-skilled workers. In this case, the technology helped less-experienced workers gain access to some of the practices used by stronger performers. AI therefore acted partly as a knowledge-transfer and learning mechanism.

From an HRM perspective, this is important because technology may influence not only productivity but also learning curves, training needs, job design and the distribution of organisational knowledge.

But Job Redesign Can Also Make Work Worse

It would be a mistake, however, to assume that removing repetitive tasks automatically improves work.

The OECD (2023) reports that many workers using AI experienced improvements in job enjoyment, autonomy and decision support. In its surveys, 63% of AI users in finance and manufacturing reported that AI had improved their enjoyment of work to some degree. AI was also often associated with fewer repetitive administrative tasks.

Yet the same OECD evidence identifies a serious risk: work intensification. Among AI users surveyed, 75% in finance and 77% in manufacturing reported that AI had increased the pace at which they performed their tasks. OECD case studies also found that when organisations automate the “easy” parts of a job, employees can be left with a continuous stream of difficult work.

The Job-Design Paradox

Removing boring tasks may enrich a job—but removing every simple task may also remove natural mental breaks. A more “advanced” job can therefore become more exhausting if workload, autonomy and recovery are ignored.

This connects strongly with HRM and the design of work. Marchington and Wilkinson (2020) emphasise that work design influences employee experience, motivation and organisational effectiveness. Job redesign should therefore consider not only which tasks technology can perform, but also what combination of tasks creates sustainable and meaningful work for employees.

The Labour Market Will Experience Both Creation and Displacement

At economy level, technological change does not produce only one direction of movement. Some jobs decline while others expand. According to the World Economic Forum’s Future of Jobs Report 2025, structural changes across technology, demographics, economic conditions and the green transition are expected to create approximately 170 million jobs by 2030 while displacing approximately 92 million, producing a projected net increase of around 78 million jobs.

Figure 3. Projected global labour-market transformation to 2030. Author’s illustration based on World Economic Forum (2025).

These projections are not a forecast of AI alone; they combine several major forces affecting the labour market. Nevertheless, AI is an important part of the transformation. The report states that half of surveyed employers expected to reorient their business in response to AI, while 40% anticipated reducing staff where AI can automate tasks.

At the same time, employers expect technology-related roles such as AI and machine-learning specialists, big-data specialists and software developers to grow rapidly. Care, education, construction, delivery and other frontline occupations are also expected to expand in absolute numbers.

This demonstrates why HR professionals should avoid treating “the future of jobs” as a single organisational event. Workforce transformation involves declining tasks, growing tasks, new roles, redesigned roles, redeployment and reskilling happening simultaneously.

Reskilling Is an HR Strategy, Not Just a Training Activity

If jobs are being redesigned rather than simply disappearing, Learning and Development becomes strategically important.

The World Economic Forum (2025) estimates that if the global workforce were represented by 100 people, 59 would require training by 2030. Employers in the survey also identified skills gaps as the biggest barrier to business transformation.

Kew and Stredwick (2016) explain that HRM must respond to changes in the external business environment by developing organisational capability. In an AI context, this means that organisations should not wait until employees’ skills become obsolete before responding. Workforce planning should identify which tasks are changing and what employees need to learn before the redesigned roles fully emerge.

Effective reskilling should include more than technical AI knowledge. Employees may need:

• AI and digital literacy;
• critical thinking and verification skills;
• problem-solving;
• communication and collaboration;
• data interpretation;
• ethical judgement;
• adaptability and lifelong learning.

This combination is important because employees need to understand both how to use AI and when not to trust it.

Employee Voice and Change Management

Job redesign can fail even when the technology itself works.

Employees may fear redundancy, loss of status, increased monitoring or being expected to produce more work with fewer resources. If management introduces AI without explaining why roles are changing, resistance may be interpreted as a “negative attitude toward technology” when employees may actually be responding to uncertainty and poor communication.

The module’s emphasis on employee relations, organisational power, voice and change management is therefore directly relevant. Employees should have opportunities to contribute to decisions about how their work will be redesigned. They possess practical knowledge about which tasks are frustrating, which tasks require judgement and where technology creates new problems.

Meaningful employee involvement can also improve implementation. A job redesigned only from management data may look efficient on a dashboard but work poorly in practice.

The people who perform a job should have a voice in how that job is redesigned.

A Global HRM Perspective

The effects of AI will also vary significantly between countries. Gmyrek et al. (2025) show that occupational exposure differs across income groups because economies have different occupational structures and levels of digitalisation. The highest-exposure jobs are more common in high-income countries, while lower-income economies currently have a larger share of work with limited exposure.

However, lower exposure should not automatically be viewed as an advantage. Countries and workers with weaker digital infrastructure may also have less access to the productivity and learning opportunities created by AI. This creates a possible AI divide where some workers benefit from augmentation while others are excluded from the tools and training required to participate.

International HRM therefore needs to consider not only automation risk but also access to technology, education, language, digital literacy and organisational support. A multinational organisation cannot assume that the same job-redesign strategy will work equally well in every location.

🎥 Recommended Video

World Economic Forum – The Future of Jobs Report 2025

This short video summarises the major job-creation, displacement and skills trends expected to reshape the global workforce.

▶ Watch on YouTube

My Reflection

Before examining the evidence, I tended to think about AI mainly in terms of job replacement: either a person keeps their job or technology takes it. I now see that this is too simple.

A job can continue to exist while changing significantly. Employees may perform fewer routine tasks, receive AI support for complex tasks, take on new responsibilities or work at a faster pace. This means that the quality of job redesign is just as important as the number of jobs that remain.

The customer-support study was particularly interesting to me because it shows that AI can transfer knowledge and help less-experienced employees improve more quickly. This suggests that AI could become part of Learning and Development rather than simply an automation tool.

At the same time, the OECD evidence on work intensity shows why HR must remain involved. If AI only allows organisations to increase targets and remove natural breaks from work, higher productivity may come at the cost of employee wellbeing.

Therefore, I believe the future of work should not be approached as Human versus AI. The more useful question is how organisations can create Human + AI jobs that combine technological capability with human judgement, creativity, empathy and responsibility.

AI may automate tasks. HR determines whether the result becomes job destruction, job enrichment or simply more intense work.

💬 Join the Discussion

If AI could automate 30% of the tasks in your current job, what should your employer do with the time that becomes available?

Should employees be given more meaningful responsibilities, shorter working time, new performance targets, or completely redesigned roles? Share your view in the comments.

References

Brynjolfsson, E., Li, D. and Raymond, L. (2025) ‘Generative AI at work’, The Quarterly Journal of Economics, 140(2), pp. 889–942. doi:10.1093/qje/qjae044. Available at: https://academic.oup.com/qje/article/140/2/889/7990658 (Accessed: 10 August 2026).

Gmyrek, P., Berg, J., Kamiński, K., Konopczyński, F., Ładna, A., Nafradi, B., Rosłaniec, K. and Troszyński, M. (2025) Generative AI and Jobs: A Refined Global Index of Occupational Exposure. ILO Working Paper 140. Geneva: International Labour Organization. Available at: https://www.ilo.org/publications/generative-ai-and-jobs-refined-global-index-occupational-exposure (Accessed: 10 August 2026).

Kew, J. and Stredwick, J. (2016) Human Resource Management in a Business Context. 3rd edn. London: CIPD.

Marchington, M. and Wilkinson, A. (2020) Human Resource Management at Work. 7th edn. London: CIPD.

OECD (2023) ‘Artificial intelligence, job quality and inclusiveness’, in OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market. Paris: OECD Publishing. Available at: OECD Employment Outlook 2023 (Accessed: 10 August 2026).

World Economic Forum (2025) The Future of Jobs Report 2025. Geneva: World Economic Forum. Available at: https://www.weforum.org/publications/the-future-of-jobs-report-2025/ (Accessed: 10 August 2026).