Barclays plans 5,000 job reductions despite rising profits. The deeper concern is whether banks are eliminating the junior positions that prepare their future specialists.

On 22 October 2026, Barclays reported £2.3 billion in quarterly pre-tax profit, an increase of 12% from a year earlier. The British bank also disclosed plans to reduce its workforce by around 5,000 positions during 2026.

About 1,000 jobs had already gone. Another 4,000 were expected to disappear before December ended.

The bank was making money, yet thousands of employees were losing their positions. That contrast made headlines. What interested me more was the kind of work being affected.

According to Reuters, about 60% of the planned reductions involve support functions, including software development. Barclays has linked the restructuring to technology investment and operational efficiency, with artificial intelligence becoming part of that process.

A bank can improve its financial results while employing fewer people. There is nothing particularly new about that. The more troubling possibility is that it may also reduce the number of positions through which young employees learn their profession.

A junior banking job is not simply a salary. It is often the beginning of an education that no university can fully provide.

Barclays Is Profitable. Why Reduce Employment?

Barclays’ October results showed rising profits, although the quarterly figure still fell short of analysts’ expectations.

The workforce reductions should not all be described as employees directly replaced by AI. Some positions are disappearing through organisational changes, while others reflect investments in software and improvements to existing processes.

The distinction matters because banks have been automating operations for decades. Payment processing, account reconciliation and transaction monitoring all rely heavily on technology. New AI systems can extend that automation into work involving language and document analysis.

An employee who once reviewed a case file and prepared a routine investigation summary may now receive an AI-generated draft. A developer may use software that produces substantial amounts of code from instructions. Customer service teams can increasingly rely on systems that interpret enquiries and suggest responses.

Employees must still check the results, particularly when financial or regulatory risks are involved. But the amount of routine work assigned to each person may decline.

For management, higher productivity creates an opportunity to reduce recurring costs. It can also encourage a bank to reconsider how many junior employees it recruits.

The immediate savings are easy to calculate. The longer-term consequences for professional training are less obvious.

Young Workers Are Already Facing a Different Labour Market

Research published by Stanford University’s Digital Economy Lab in August 2026 offers an early indication of how AI exposure may affect employment. Using payroll data covering millions of American workers, researchers examined employment patterns across occupations with different levels of exposure to generative AI.

They found that employment among workers aged 22 to 25 in highly exposed occupations was about 19% below the level suggested by the employment trend among similarly aged workers in less-exposed occupations. The study found no comparable gap among experienced workers.

Much of the adjustment appeared to occur through weaker hiring rather than unusually high rates of dismissal.

The researchers did not claim that AI caused the entire difference. Other changes in the labour market may have contributed, and the findings cannot automatically be applied to every country or profession.

Still, the age difference deserves attention. A graduate entering a bank usually starts with tasks that follow established procedures. Preparing documents, checking transactions and responding to routine enquiries provide opportunities to understand the institution.

An experienced employee approaches the same work differently. Years of dealing with exceptions help that person recognise problems that are not immediately apparent from the available information.

AI can reproduce parts of a standard workflow. The judgment developed through difficult cases is harder to replace.

If banks increasingly rely on experienced employees supervising automated systems, younger workers may struggle to secure the positions that would allow them to acquire similar experience.

What Working With Cross-Border Payments Taught Me

I spent much of my professional life working with cross-border payment systems. That background shapes how I understand the employment debate surrounding AI.

A payment instruction can look perfectly routine until the receiving bank reports that the funds have not arrived. Sometimes the problem lies in incomplete message details. On other occasions, a correspondent bank needs clarification before processing can continue. A compliance alert may complicate an otherwise straightforward transaction.

Each case requires an understanding of payment procedures, the responsibilities of the banks involved and the information available to resolve the problem.

Written procedures remain essential. They provide consistency and help employees meet operational requirements. Yet a procedure cannot anticipate every unusual situation.

Working with cross-border payments taught me that knowing the required steps is not always the same as understanding the problem.

An employee may follow a checklist correctly and still struggle when an investigation produces conflicting information. Experience becomes valuable when the obvious explanation fails.

Much of that knowledge develops through repeated exposure to difficult cases, often with guidance from more experienced colleagues.

AI can help identify missing information and prepare investigation summaries. It may also detect patterns that a human employee would overlook.

These capabilities are valuable. Faster investigations can reduce delays for customers and improve the bank’s efficiency. The role of international payment infrastructure is another reminder that financial technology has strategic as well as operational consequences.

But financial institutions remain accountable for their decisions. A sanctions alert, for example, cannot simply be dismissed because a system produces a reassuring explanation. The bank must follow the applicable rules and maintain effective controls.

Human judgment therefore remains important, although the skills required to exercise it will change.

The training problem arises when automation removes much of the routine work that once helped employees develop that judgment.

A bank may retain its experienced specialists while reducing junior recruitment. For several years, the arrangement could work perfectly well. The difficulties may appear later, when senior employees retire or leave and fewer trained replacements are available. This question also connects with the changing meaning of work and retirement in an AI economy.

The Pressure Extends Beyond Barclays

Barclays is not alone in examining how technology can reduce operating costs. Standard Chartered has outlined plans to eliminate more than 7,000 corporate-function positions by 2030. The programme includes automation and organisational restructuring.

Other large employers are also reconsidering the staffing requirements of work that software can increasingly perform. The consequences will vary considerably across industries.

London has a large concentration of financial services and professional occupations whose tasks are exposed to AI. Yet exposure is not the same as displacement. Some employees may become more productive without losing their jobs, while others may find their responsibilities changing substantially.

The international effects deserve attention as well. For years, banks and multinational companies have moved technology and administrative functions to countries where they can recruit skilled workers at competitive costs.

India has become a major centre for these services. Pakistan’s technology and business-process outsourcing industries operate in parts of the same global market.

AI could reduce demand for some routine outsourced tasks. At the same time, companies capable of managing complex systems or providing specialist services may discover new opportunities.

The outcome will depend on investment decisions and the ability of employees to acquire new skills. A country cannot assume that demand for low-cost digital labour will continue indefinitely simply because it has a large pool of educated workers.

Economic Growth Does Not Guarantee More Jobs

The Barclays announcement raises a question that extends beyond banking. A business can increase output without hiring additional employees. It may even expand while reducing its workforce.

That does not mean economic growth has stopped creating jobs across the economy.

Technological change has repeatedly altered employment patterns. Banks once required large numbers of employees to perform calculations and process paper records. Computers removed much of that manual work, while expanding financial networks created other occupations.

The introduction of ATMs offers a familiar example. Machines reduced the labour required for many cash transactions, but their broader employment effects depended on how banks changed their branch networks and services.

AI may produce a comparable transformation, although its effects could differ because it can perform some tasks traditionally associated with professional knowledge.

The speed of adoption is especially important. Workers displaced from one occupation cannot always move immediately into another. New positions may require experience or specialist knowledge that existing employees lack.

Young graduates face an additional difficulty. They may never enter the occupations in which earlier generations acquired their professional skills.

The distinction between productivity growth and employment growth has consequences for public policy. As I examined in an earlier article, Europe’s economic growth does not automatically make workers feel better off.

Governments depend on employment to support household income and tax revenue. If a larger share of economic gains flows to capital while wage growth weakens, the distribution of those gains becomes increasingly important.

The challenge is not necessarily an economy without workers. It may be an economy in which productivity improves while secure career opportunities become harder to obtain.

Tax Policy May Influence the Choice Between Workers and Technology

Companies decide whether to invest in software partly by comparing expected productivity gains with the costs of employing people. Tax rules can influence that calculation.

In June 2024, the International Monetary Fund warned that favourable tax treatment for some software and equipment investments could encourage businesses to replace labour.

The IMF did not recommend a simple tax on robots. Such a measure could discourage innovation, including technology that helps employees perform their jobs more effectively.

Instead, it urged governments to reconsider tax incentives that favour labour-replacing investment and to strengthen support for workers affected by technological change.

The distinction is important. Replacing a repetitive task with software may improve productivity and benefit customers. Penalising every form of automation would make little economic sense.

But governments should examine whether their tax systems encourage companies to replace employees even when the wider social costs are substantial.

Public finances also face a distribution problem. When a company increases profits through automation, the tax system must determine how the resulting gains contribute to public revenue. Workers who lose their jobs may need retraining or temporary income support.

Such costs do not necessarily appear in the company’s calculation of its return on investment. A policy that rewards technological efficiency without considering workforce adjustment risks transferring part of the cost to society.

Banks Must Rethink How They Train Employees

Preventing banks from adopting AI would be neither practical nor desirable. Customers expect payments to move quickly. They want errors corrected without unnecessary delays, and modern financial institutions must process vast amounts of information while meeting demanding regulatory requirements.

The challenge is to preserve professional learning while changing the way work gets done.

Banks could redesign junior positions so that employees learn to investigate complex cases under supervision, rather than spending years completing tasks that software can already manage.

Training may need to place greater emphasis on understanding systems, interpreting unusual transactions and recognising the limits of automated decisions.

Experience will still matter. Its development may require more deliberate planning than it did when junior staff learned through large volumes of routine operational work.

That has implications for recruitment budgets. A bank can calculate the immediate savings from eliminating several hundred positions. It is harder to estimate the value of an employee who might become a specialist ten years later.

Managers working under pressure to meet annual financial targets may naturally concentrate on the first figure.

Yet banking institutions are expected to remain reliable for decades. Their workforce planning should reflect that longer horizon.

The Cost That May Not Appear in Barclays’ Accounts

Barclays’ restructuring illustrates how modern banks can pursue higher productivity while reducing employment. The announcement does not prove that AI has destroyed the relationship between economic growth and jobs. Nor does it establish that every position removed by Barclays was replaced by technology.

It does, however, bring attention to a problem that deserves more consideration.

Experienced employees are valuable partly because of work they performed years earlier, often in positions that now appear suitable for automation.

The knowledge accumulated during those early years cannot always be recreated through a training course or an AI-generated explanation.

Barclays can measure the savings from reducing its workforce. Measuring the loss of future institutional knowledge will be more difficult, particularly while experienced employees remain available to resolve complicated cases.

The shortage may become apparent only when the next generation is expected to take their place.