For much of 2026, the most dramatic predictions about Artificial Intelligence and employment have appeared to be wrong.
There has been no sudden collapse in professional employment, offices remain busy and most people working in finance, accountancy and administration are still performing broadly the same jobs. This has encouraged some commentators and employees to dismiss the warnings and conclude that AI has been considerably overhyped.
However, the absence of mass redundancies does not mean that nothing is changing.
A significant indication of what may be coming can now be found in graduate recruitment. UBS is reportedly requiring prospective graduates and interns joining its global banking and markets divisions in 2027 to demonstrate how they can use AI to improve efficiency and produce better results.
According to the Financial Times, the new requirement will sit alongside traditional academic qualifications. AI-related questions are also expected to be included in recruitment interviews.
UBS already promotes AI prominently on its early careers website, telling prospective recruits that they will receive hands-on learning, access to internal AI certifications and opportunities to participate in global knowledge-sharing communities.
This is an important development because it suggests that AI proficiency is moving beyond specialist technology positions. It is becoming a skill expected from people applying for traditional banking and financial services roles.
The Entry Requirements Are Changing
According to a Financial Times report, the new requirement will sit alongside traditional academic qualifications. AI-related questions are also expected to be included in recruitment interviews.
A graduate might previously have distinguished themselves through a strong degree, an internship, knowledge of financial markets and good analytical ability. Those qualities remain important, but employers may increasingly expect candidates to demonstrate that they can also use AI competently and responsibly.
That does not simply mean asking a chatbot to write an email.
In a financial services environment, employees may need to know how to use AI to research companies, analyse information, summarise documents, prepare initial reports and improve presentations. They must also understand the limitations of the technology, including inaccurate outputs, bias, confidentiality risks and the continuing need for human oversight.
The successful candidate of the near future may therefore be the person who combines financial knowledge and professional judgement with an ability to use AI effectively.
That creates a new risk for graduates and existing employees who have dismissed the technology because it has not yet caused a visible upheaval in their own workplace. They may discover that employers have quietly changed what they expect while the wider debate remained focused on whether AI had produced mass unemployment.
The Calm Before the Storm?
We recently examined this issue in our article, AI Has Not Reached Your Office Yet, But Is It Being Developed Elsewhere in the Network?.
The article argued that employees should not assume their organisation is doing little with AI simply because they have not seen substantial changes in their own office.
A bank, insurer or international accountancy practice does not need every branch or jurisdiction to develop its own AI systems. Specialist teams can build and test the technology in London, New York, Singapore or another centre before distributing it across the organisation.
The process of development may take years, particularly in regulated industries where security, confidentiality, accuracy and governance must be carefully addressed. Once a system has been tested and approved, however, its introduction across an international network could happen much more quickly.
This may help explain why 2026 has sometimes felt like the calm before the storm. Much of the important activity is taking place behind the scenes through testing, staff training, data preparation and the development of internal controls.
The UBS recruitment requirement supports this argument. The bank is preparing its next generation of employees for a workplace in which AI is expected to play a much larger role.
Job Losses Are Not the Only Measure
The latest evidence does not show that AI is already eliminating vast numbers of jobs.
A September 2026 survey from the Federal Reserve Bank of New York found that AI use had risen sharply among businesses in its region. Sixty-one per cent of service firms reported using AI, compared with 40 per cent in 2025 and 25 per cent in 2024. Usage was particularly high in knowledge-intensive sectors, including finance, business services and information.
Nevertheless, only 4 per cent of AI-using service firms said they had made redundancies because of the technology during the previous six months.
At first glance, that appears reassuring. A closer examination presents a more complicated picture.
Approximately 15 per cent of the service firms said they had recruited fewer employees than they otherwise would have because of AI. At the same time, 13 per cent had hired additional people to help them use the technology, while more than a third were retraining existing employees.
AI is therefore not producing one simple employment outcome. It can reduce recruitment in one area, create demand elsewhere and change the skills required within jobs that continue to exist.
This is why counting announced redundancies provides only part of the picture.
A company may not dismiss hundreds of employees. It may allow its workforce to decline gradually as people leave, recruit fewer graduates or expect the same department to manage more work without increasing its headcount.
The job titles may remain, but there could eventually be fewer positions available and greater competition for those that survive.
Junior Employees Could Feel the Change First
Entry-level work is particularly exposed because junior employees have traditionally undertaken many of the tasks AI can now assist with.
These include preliminary research, document review, data extraction, routine analysis, report preparation and the production of presentations. AI may not be capable of replacing an experienced banker or investment professional, but it can reduce the amount of supporting work required before an experienced person makes the final decision.
This does not necessarily mean that graduate opportunities will disappear. It could mean that employers recruit fewer people and place greater value on applicants who can supervise, verify and improve AI-generated work.
The danger is that young people may be told there is no need to prepare because the much-publicised AI jobs apocalypse has not occurred. By the time the effects become obvious, employers may already be favouring candidates who began developing these skills several years earlier.
Neither Panic Nor Complacency
Claims that AI will immediately replace entire professions remain exaggerated. Financial services still require human judgement, accountability, communication and an understanding of individual clients and circumstances.
However, dismissing AI because most jobs have survived 2026 would be equally shortsighted.
The first major change may not be a wave of redundancy announcements. It may be a new sentence in a job advertisement, an AI-related question during an interview or an expectation that existing employees use approved AI tools as part of their daily work.
UBS appears to be providing an early example of that transition.
AI may not have transformed every office yet, but it is already beginning to transform the qualifications needed to enter one.
For graduates, financial professionals and anyone preparing for the future of work, the message is increasingly clear: understanding AI is no longer merely an advantage. It is becoming part of being employable.
Prepare for the Changing Workplace
AI skills are becoming essential, whether you are entering the workplace, developing your career or preparing for changes within your profession.
AI Tuition Hub provides access to over 110 practical AI courses for just £9.95 a month. Full course descriptions and lesson titles can now be viewed before you decide to join.
Published: 6th September 2026.