The emergence and rapid development of artificial intelligence are fundamentally reshaping the world – and the role of human beings within it. AI is taking over an ever-growing range of tasks that until recently were considered uniquely human forms of intellectual work: analyzing information, identifying patterns, preparing texts, processing data, and generating ideas. Yet the paradox of technological transformation is that as machines become more capable, human qualities do not become less important. Rather, the nature of their value is changing. In many ways, AI is bringing us back to the ideal of versatility that defined the Renaissance, when people were expected to look beyond narrow specialization, constantly rebuild their skill sets, and, just as importantly, unlearn skills that have become obsolete. Continuous learning, analytical thinking, and the ability to adapt are becoming some of the defining requirements of the global labour market.
What competencies will determine professional value in the years ahead? Why does artificial intelligence not make human expertise obsolete, but instead raise the bar for it? And which skills should people be developing today? These are among the questions explored by banker, financier and Deputy Chairman of the Management Board of VTB Dmitry Breitenbikher.
Editorial translation
Photo: Vyacheslav Novikov
Until recently, a professional career was generally seen as a fairly linear progression: education, a first job, accumulated experience, and increasing responsibility. A person might change professions, but the underlying model remained the same – acquire expertise in a particular field and then convert that expertise into professional value over decades.
That model is rapidly becoming a thing of the past. According to estimates cited by the World Economic Forum, by 2030 around 39% of the key skills required in the labour market will be changed. At the same time, some 92 million jobs could disappear, while around 170 million new ones could be created. The result is not so much a decline in employment as a large-scale restructuring of the labour market. Some 59% of workers are expected to require retraining.
This is an important nuance in the debate about artificial intelligence. The key question for the future of work is not how many professions technology will eliminate. A much more important question is: what kind of person will remain in demand in an economy where a significant share of intellectual routine has already been automated?
The answer is gradually becoming clear. Value is shifting away from what a person knows towards how they think, how they work with technology, and how quickly they can reinvent their professional model.
In the new economy, it is no longer enough to simply be a good specialist. This does not mean that professional expertise is becoming irrelevant. On the contrary, deep subject-matter knowledge remains the foundation. But expertise alone is no longer enough to guarantee a lasting advantage.
Broadly speaking, emerging competencies can be divided into three groups.
The first is technical: analytical thinking, data literacy, artificial intelligence, and systems thinking. What distinguishes these skills is that the tools within this category are evolving rapidly.
The second is human: leadership, communication, and the ability to navigate change. These skills, by contrast, are becoming more valuable precisely because they are much harder to automate.
The third is meta-skills: the ability to learn, creativity, and interdisciplinary thinking. They determine how quickly a person can acquire everything else.
The formula for future professional resilience is therefore relatively simple: technical skills need to be constantly updated, human skills need to be deepened, while meta-skills become the multiplier across the entire system. This combination is emerging as the defining form of human capital in the age of AI.
It is telling that analytical thinking now ranks among the most sought-after skills by employers. The reason is obvious: the amount of information available to us is growing faster than our ability to make sense of it. Artificial intelligence can already search for information, write texts, generate images, and produce computer code. But having an answer does not necessarily mean having the right solution. A human being is still needed to formulate the problem, develop a hypothesis, assess the evidence, distinguish correlation from causation, and, above all, ask the question: is the answer actually correct? The more decisions are delegated to machines, the greater the premium on human judgement. In this sense, analytical thinking is no longer simply the ability to calculate. It is the ability to question the first obvious answer.
This changes the very logic of professional advantage. If experience was once the primary source of efficiency, some of that accumulated experience can now be embedded in technology. The advantage increasingly goes to those who can formulate a task correctly for AI, critically evaluate the result, and integrate it into their workflow. The key question, therefore, is no longer “Will AI replace humans?” but rather: “What kind of human will AI make stronger?”
The history of Capital One illustrates this logic well. The bank turned credit-card operations into a continuous process of experimentation, testing different interest rates, credit limits, and ways of communicating with customers. Instead of relying on managerial “we think”, it developed a culture of “we tested it”. Analytics became not a supporting tool but part of the business model itself.
In the age of AI, this principle takes on a new significance: a machine can dramatically accelerate analysis, but a human being remains responsible for asking the right question and interpreting the result.
The fastest-growing skills are increasingly related to AI and data. But it is particularly important to avoid a common misconception: AI literacy cannot be reduced to knowing how to formulate the right prompt for a chatbot. True AI literacy means understanding both the capabilities and limitations of a model: what it does well, where it can make mistakes, how to verify its output, what data can safely be shared with it, and what risks its use may create.
An experiment conducted by Erik Brynjolfsson and his colleagues is particularly revealing. The use of an AI assistant in a customer-support operation increased productivity by approximately 14%, with the greatest gains recorded among less experienced employees. In this case, the technology did not simply make work faster – it narrowed the gap between newcomers and more experienced specialists.
The more complex the economy becomes, the less effective solutions are when they treat problems in isolation. Supply chains, climate risks, digital transformation, and artificial intelligence are all systems characterized by multiple feedback loops. A change in one element can trigger consequences far beyond the original problem. This is why systems thinking is becoming increasingly important: the ability to see not isolated events but the relationships between them, to understand delays and unintended consequences, and to identify points at which a small intervention can alter the behaviour of an entire system.
The history of Kodak has become a textbook example of what happens when this perspective is missing. The company itself developed one of the first prototypes of a digital camera, yet digital technology was perceived as a threat to its existing film business. At the level of an individual decision, protecting a successful business model appeared rational. At the level of the entire system, it proved to be a strategic trap.
The paradox is that a company can possess the technology it needs and still lack the ability to see how a change in one element will reshape the entire market. Systems thinking makes it possible to move from asking “What is happening?” to asking “What will this decision set in motion next?”
Data are often described as the new oil. For the modern economy, however, it may be more accurate to think of them as fuel: they create value only when they are transformed into decisions. A company can possess terabytes of information and have no idea what to do with it. Conversely, a single correctly interpreted dataset can change a company’s strategy. Working with data therefore requires not only technical skills but also critical thinking. Where did the data come from? Who is missing from the dataset? What limitations are embedded in it? Is the relationship we have identified causal, or is it simply a coincidence?
The story of US retailer Target illustrates another dimension of the issue. By analyzing purchasing patterns, the company was able to identify with a high degree of probability which customers were pregnant, sometimes before the information had become obvious even to their relatives. From an analytical perspective, this was an impressive achievement. From a business and societal perspective, however, it raises a very different question: where is the line between the ability to use data and an individual’s right to privacy? The more powerful analytical tools become, the more important the ethics of their use becomes.
At this point, it becomes clear why human skills do not disappear with the spread of AI but, on the contrary, become more valuable. Leadership is emerging as one of the fastest-growing skill areas. This is logical: when part of intellectual work is performed by machines, human leaders must focus on what cannot be fully delegated to an algorithm – creating meaning, coordinating people, building trust and taking responsibility.
The teams of the future will increasingly be built around a “human plus AI” model. In such an environment, leadership is no longer simply about allocating tasks. It is about creating a setting in which people do not compete with technology to prove their own value, but use it to improve collective performance. Psychological safety takes on particular importance: the ability to speak openly about mistakes, ask questions and challenge decisions without fear of being punished simply for disagreeing. This may sound like a soft management issue, but it has a direct economic impact. A team in which people are afraid to report a problem stops learning. And a team that stops learning will inevitably fall behind.
If almost 40% of key skills are changing within a few years, transformation can no longer be treated as a standalone project. It becomes a permanent state of the organization. A new CRM system, digital platform, algorithm or organizational structure may be technically flawless – and still fail to deliver results. The reason is simple: people change their behaviour only when they understand why change is necessary, are willing to embrace it, and are given the opportunity to learn a new way of working. Technology implementation is therefore always both a technological and a human challenge.
The difference between a successful and an unsuccessful transformation often has little to do with the quality of the software itself. In one case, employees are involved from the outset, the purpose of the changes is explained, and they are given an opportunity to master the new tool. In another, a new way of working is simply announced. Formally, the system has been implemented. In practice, employees continue to work as before. Technological transformation without a transformation in human behaviour is ultimately little more than the installation of new software.
In an age of information overload, communication becomes one of the main force multipliers of every other skill. A brilliant analysis may be produced, but if it cannot be explained to the person making the decision, it creates little value. A powerful strategy may be developed, but if the team does not understand its purpose, the strategy will remain just another document. Communication, in this sense, is not the ability to speak at length or eloquently. It is the ability to reduce complexity without losing meaning.
Steve Jobs’s presentation of the first iPhone offers a powerful illustration. Rather than simply listing technical specifications, he presented the product through a simple concept: three different devices that were, in fact, one. First came the meaning, then the emotional connection, and only afterwards did the technical specifications serve as evidence. In a world where AI can produce an almost limitless stream of content, the advantage will not go to those who say more. It will go to those who can identify and articulate what matters most.
There is another myth that needs to be reconsidered: that creativity is an innate gift. In reality, creativity can be trained. It arises not only from inspiration but from the ability to generate multiple options, connect previously unrelated ideas and test hypotheses quickly. This is where artificial intelligence can become not a competitor but a partner. If machines take over the routine part of the work, people gain more space to ask new questions and create new combinations.
The history of Netflix is instructive for precisely this reason. The company did not attempt to defend indefinitely the business model that had made it successful. It began investing in streaming and later in original content, effectively reinventing its business model several times. Creativity in this sense is not the ability to come up with one brilliant idea. It is the willingness to continually question a model that is already successful.
This changes the very logic of professional advantage. If experience was once the primary source of efficiency, some of that accumulated experience can now be embedded in technology. The advantage increasingly goes to those who can formulate a task correctly for AI, critically evaluate the result, and integrate it into their workflow. The key question, therefore, is no longer “Will AI replace humans?” but rather: “What kind of human will AI make stronger?”
The more complex a problem becomes, the less likely it is that it can be solved within a single discipline. A financial specialist who understands only finance, a marketer who understands only marketing, or a technologist who does not understand user behaviour is constrained by their own professional lens. Increasingly valuable are T-shaped professionals – people with deep expertise in one field combined with a broad understanding of others. New solutions often emerge precisely at the intersection of disciplines: finance meets psychology, medicine meets data, engineering meets design.
The research of Scott E. Page on the diversity of models of thinking provides an important intellectual framework here. In complex problem-solving, diversity of perspectives can give a team an advantage over a group of people who share the same professional logic, even when each individual member of that group is exceptionally capable. In other words, a complex world does not require one perfect way of thinking. It requires a set of different lenses.
All the skills discussed above share one common foundation: the ability to learn. If almost half of the skills people need are changing within a few years, it is impossible to build a career solely on accumulated knowledge. People must not only continuously acquire new skills but also let go of what worked until recently.
This is a crucial point. The ability to learn is not only the ability to acquire something new. It is also the ability to unlearn. Experience can be an advantage, but it can also become a trap. The longer a person works according to a particular model, the greater the risk of confusing what is familiar with what is right. Professional resilience today is therefore determined not by the volume of accumulated experience, but by the speed at which that experience can be renewed. Research on learning shows that simply rereading material creates more of an illusion of knowledge than durable learning. Active recall, spaced repetition and practice with feedback are considerably more effective.
The same logic applies far beyond education. Anyone who wants to remain professionally competitive must regularly put themselves in situations where there is something they do not yet know how to do.
There is an understandable temptation to create a ranking of future competencies and then acquire them one by one. But this is precisely where the main mistake lies. In real-world work, skills create value not individually but in combination. Analytical thinking combined with data creates the ability to draw causal conclusions. The ability to learn combined with AI makes it possible to master new tools in days rather than months. Leadership combined with communication creates the trust needed to drive change. Creativity combined with interdisciplinarity makes it possible to find solutions beyond the boundaries of an established professional model.
Competencies therefore make more sense when developed not as a collection of certificates but as a system of mutually reinforcing capabilities. This leads to a practical principle: it is better to develop two skills simultaneously through a real-world challenge than to learn ten skills sequentially and in the abstract.
If everything discussed above were reduced to a single formula, it would look something like this: the future belongs not to the narrowest specialisation, but to a combination of thinking, AI literacy, and human skills. This does not mean that experts will disappear. On the contrary, deep expertise will remain essential. But its role will change.
The specialist of the future will not simply need to know more than others. They will need to combine deep knowledge with analytics, technology, communication, and continuous learning. In this sense, artificial intelligence does not eliminate human value. It forces us to redefine its boundaries.
Machines are becoming increasingly effective at searching, processing and generating information. As a result, information itself is becoming cheaper. What becomes more valuable is the ability to determine which information matters, which question is actually worth asking, which decision should be made and how to persuade others to act.
This is why the key competitive advantage of a human being is no longer the amount of knowledge accumulated by a certain age, but the speed at which they can acquire new knowledge, abandon outdated models and connect different fields into a new picture. A profession is ceasing to be the final destination of education. It is becoming a temporary configuration of skills. And that means the most promising strategy for the coming decades is not to guess which profession will be in demand in 2050. That is practically impossible. A far more rational strategy is to develop the ability to remain in demand, whatever shape the world of work may take.
