The human touch: How analogue skills provide job security in a digital world

1. This is the number of jobs that computer science graduate Zach has unsuccessfully applied for since completing his bachelor's degree in 2023. In an interview with the New York Times last month, he reports on the perseverance it takes to continue applying for open positions. Yet Zach is not alone in his predicament. According to a Stanford study from August 2025, there has been a 13% decline in employment of 22–25-year-old graduates in the USA since 2022.

Here in Germany, the situation looks similar: over the last five years, we have witnessed a 66% increase in the unemployment rate among young graduates. According to the Federal Employment Agency, this rate alone rose by 25% in the last year. In absolute figures, that means: in summer 2025, Germany had approximately 44,000 unemployed graduates under the age of 30. These statistics should send a shiver down our spines as students. The bubble of educational illusion has burst. Where five years ago there was talk of job security as long as you studied – especially with a degree in STEM – today heads are shaken in despair.

In fact, the unemployment rate among graduates from precisely these subject areas is the highest. Yet no industry seems safe, whether in finance, IT, pharmaceuticals, marketing or law: the unemployment rate among entry-level professionals is rising rapidly worldwide. In the English-speaking world, people are now talking about the so-called "Graduate Job Crisis". The root of all evil? AI.

Since the widespread adoption of chatbots like ChatGPT, companies have been toying with the idea of replacing tasks that fall within the remit of entry-level professionals with AI. It is cheaper, more effective, faster and saves staff the effort of training new colleagues. A large number of companies have now begun to develop and deploy internal AI tools. These can take on a range of tasks: data analysis and evaluation, software programming, customer service, retrieval and application of existing knowledge, routine writing, administration, translation, visual design, etc. Tasks that in the past were often carried out by junior staff.

The international law firm Simmons+Simmons has been using a GenAI tool called "Percy" since February this year. This takes over all research, analysis and administrative tasks of junior lawyers and legal assistants.

Other firms are not quite there yet, but they too are not hiring new entry-level staff. Experts are calling out a standstill in the global job market. In this time of AI revolution, there is great uncertainty on the market. Companies do not know how and to what extent it makes sense to incorporate artificial intelligence into their business model. But as long as no decision has been made, no new employees will be hired. A paralysis that, in an era characterised by buzzwords such as "dynamism" and "change", seems extremely ironic. According to the job platform Indeed, the global number of open entry-level positions alone fell by 33% in the last year. However, experts at the Financial Times warn against hastily laying off current employees and imposing a hiring freeze in favour of AI.

A central concern is the overestimation of the actual capabilities of the machines. If an AI model is to effectively replace masses of employees, it must be reliable, cybersecure, precise and fast. However, reality looks different.

Many large German corporations, including BMW, Siemens and Allianz, now use so-called "Closed AIs" internally. Here, a company purchases a licence for so-called "enterprise versions" of well-known AI models such as DeepL, ChatGPT, etc. These are then deployed on the company's private cloud environment and external access is blocked for data protection reasons. The model thus only interacts with employees and is fed exclusively with internal data. The result: significant losses in the quality and further development capability of the AI.

Large language models are based on the idea of machine learning. They continuously improve the precision of their output by being fed a constant stream of feedback from public user interactions. However, this feedback loop is cut off in such closed instances. The AI stops learning and becomes faulty.

In everyday working life, this quickly becomes apparent. "Even with simple translation tasks, the model reaches its limits. When translating a document, for example, the same word was translated differently and incorrectly within a single paragraph," says a BMW employee from Munich. "Instead of having tasks taken off our hands, we have to do double work and painstakingly improve the AI's output."

Beyond these internal company problems, replacing junior talent with AI has the potential to trigger market-wide crises. If there are to be fewer and fewer entry-level positions whilst the number of staff in middle and senior management remains the same, a simple mathematical problem emerges: the pipeline no longer works. Without sufficient qualified junior staff, these positions will not be able to be filled in the future. Either completely untrained, unqualified employees will be hired, or companies will continue to shrink.

Unfortunately, a large proportion of companies are still closing their eyes to these long-term effects. So what can we aspiring graduates do now to start a future-proof career? Where should we invest time and effort to optimise our job prospects? Economists at Harvard University declare: "The age of specialists is over."

We must realise that in every niche, expert knowledge is being replaced by AI and its data collection and analysis capabilities ever faster. It is therefore completely pointless to play the big guessing game when starting university and bet that knowledge from a particular subject area will be in high demand in three years' time. Because within three years, AI will have learned significantly faster than we have and can replace us with our own knowledge again. The specialist race is one that we will now always lose against the machines.

And yet all is not lost. Instead of competing with AI, we should focus on what sets us apart from it: our human touch.

In plain terms: our analogue skills – or as the HR manager would put it: our soft skills. Paradoxical that it is precisely analogue skills that should provide job security in the digital age. Yet in Google's Project Oxygen, an internal performance review, it emerged that the best managers possess a set of skills that characterise their excellent work: communication of complex ideas, teamwork ability, critical engagement with language, logical thinking, emotional intelligence and flexibility. Skills that are essential for the survival of increasingly digitalised companies.

The ability to adapt to constantly changing intelligent tools, to describe complex problems to them clearly and comprehensibly, to critically engage with the outputs of language models and to apply these insights meaningfully in an international, interdisciplinary context will become indispensable in the working world of the future. By improving these qualifications, "a workforce is created that is resilient, can adapt quickly and is ready to succeed in a constantly evolving environment," according to an article from August 2025 in the Harvard Business Review.

Instead of rigidly pursuing specialist status, we should focus on developing these skills. Only in this way can we emerge as winners from this AI revolution.

Here then is my official appeal to the German education system: it is high time for a systemic change. Time for an education system in which value is placed on the acquisition of these skills from the outset. Other global players have a head start on us in this regard. Let us look at Singapore, for example, a country known as a leader and early adopter in digital transformation. There, "social-emotional learning" and "21st-century competencies" are concrete frameworks for learning soft skills in the national curriculum. Through these, communication, collaboration and self-management are emphasised from primary school through to higher education. In Finland and Australia too, these analogue skills have been taught as a fixed part of the national curriculum since the early 2010s. Let us follow the example of these countries (incidentally, both in the top 10 of the world's happiest countries in 2024) and reflect on what makes us so special.

The path to success is therefore to place personal and professional focus on that certain "je ne sais quoi" that we all possess. On our human touch.