Commentary: AI in the Workplace We're Paying to Make Ourselves Obsolete

Related Vendor

Taiwan Machine Tool & Accessory Builders' Association (TMBA)

We have created a digital world. Now we are creating a monster within it. That sounds drastic, perhaps even exaggerated. But the longer I observe the development of artificial intelligence, the harder it is for me to find a less alarming way to describe it.

In his blog, Bill Gates warns that AI could become either the greatest equalizer in history or a massive source of new inequality. I fear that we are currently heading toward the second possibility with astonishing consistency. Of course, I see the advantages. In medicine, artificial intelligence can analyze vast amounts of data and recognize patterns that a human might miss. It can compare image data, identify abnormalities, and assist doctors with diagnoses. The same applies to research, industrial processes, and many other fields. AI is an extraordinarily powerful tool.

Fears about the “end of work” have been with us since the invention of the steam engine. So far, automation has generally not eliminated jobs, but rather transformed the nature of work and created entirely new industries. However, AI differs in the speed and scope of its adoption. While earlier waves primarily affected physical labor, AI strikes at the heart of knowledge-based and creative work. The question, then, is how quickly policymakers and society will take countermeasures through regulation, new tax models (e.g., a value-added tax), and guidelines.

AI is Not an All-Knowing Inventor

The industry currently pins its hopes on AI to provide quick solutions to nearly every problem. In doing so, it’s easy to overlook what today’s AI systems are based on: existing data, existing knowledge, and the results of human work. From these, they can derive impressive correlations and generate new combinations. This is more than just looking something up. However, it is by no means a guarantee of a genuine leap in understanding.

AI can generate plausible answers based on known information. Whether these answers are correct, novel, technically feasible, or socially meaningful still requires human judgment. However, when there is insufficient data and we are truly entering uncharted territory, AI reaches its limits. Nevertheless, companies are currently cutting jobs or no longer filling them, hoping that AI will take over this work in the future. Ironically, entry-level positions are particularly affected. But this doesn’t just mean the loss of a job. It also means the loss of the place where people gain experience, make mistakes, and develop into the skilled professionals who could later evaluate the results of an AI.

Anyone who cuts entry-level positions today because AI is handling the most basic tasks shouldn’t be surprised tomorrow when they find themselves short on experienced employees. We’re not just automating work; we may also be disrupting the transfer of knowledge. Even more serious is the economic shift. Millions of people could lose their jobs or, at the very least, forfeit part of their income and bargaining power. However, the systems that are taking over their tasks do not belong to the general public. They belong to a handful of companies that control the models, data centers, platforms, and access—OpenAI, Meta, and Google, to name just a few.

Companies pay these few providers to cut back on human labor. Employees often even help train the systems through their texts, programs, images, decisions, and feedback. In the end, productivity, knowledge, and value creation flow to a handful of platform operators—who, as we all know, are almost never based in Germany. So we’re giving these companies money so that their systems can make us obsolete.

That is the real social time bomb. It’s not the machine that’s taking our jobs away of its own accord. We ourselves are orchestrating the transition, because from a business perspective, every single cost-saving measure initially seems reasonable. When a company uses AI to produce more cheaply, all its competitors come under pressure to follow suit. What makes sense for an individual company can be devastating for society as a whole.

This is because our economic system is based on people selling their labor, earning an income, and thereby participating in economic life. When labor is replaced on a large scale by systems owned by a few companies, the resulting productivity can no longer be automatically distributed through wages. As a result, profits rise while the social foundation of those profits erodes.

That is why it is not enough to constantly celebrate new AI applications and then talk about retraining for those whose jobs have disappeared. Not every laid-off administrative worker will become an AI developer. And even if everyone were properly qualified, that would not automatically create enough new jobs.

We must therefore discuss now who owns the productivity generated by AI, how it is distributed, and which tasks we deliberately want to leave to humans. Taxing AI-generated value, as Gates proposes, should not be off-limits. Nor should the question of whether certain activities should remain the domain of humans—not because a machine cannot perform them, but because we as a society do not want it to.

In the past, companies emerged from a technical idea. Engineers and developers founded them because they wanted to solve a problem, improve a product, or build something that didn’t exist before. Today, even technology-driven companies are often led by managers whose key performance indicators are not the level of innovation, technical substance, or long-term competitiveness, but rather profit margins, stock prices, and dividends. Within this logic, artificial intelligence appears to be the perfect tool for reducing the workforce.

The key difference is this: Those who approach a business from a product perspective use AI to make developers more productive. Those who think exclusively in terms of the bottom line use the same technology to cut back on developers. In the short term, this may improve the bottom line. In the long term, however, the very knowledge that the company needs to create its next products disappears from the organization. A dependency on tech giants develops. These companies can then demand whatever they want. A dividend is not innovation. Layoffs are not a strategy. And a company does not become future-proof by laying off the very people who are supposed to shape its future.

In this context, companies that convey to their employees how grateful they should be for their jobs come across as particularly deceptive. After all, this gratitude is usually surprisingly one-sided. A job is not a gift. Employees exchange their time, knowledge, creativity, and often their health for an income. They develop products, solve problems, and serve customers—and in doing so, they create the value that a company can then distribute.

If those same people can be replaced by AI, it quickly becomes clear how little the previously touted sense of community was actually worth. Then “our most valuable employees” suddenly become cost centers, and years of loyalty turn into an item on the restructuring agenda. Of course, companies must act in a business-oriented manner. But they shouldn’t talk about a sense of belonging as long as they’re turning a profit, only to invoke pure market logic the moment people seem expendable.

I notice this trend most clearly in the creative industry. Suddenly, the same flawless content is popping up everywhere. Everything is technically polished, produced quickly, and looks professional at first glance. But it’s becoming increasingly rare to see a human touch in it. At this point, I’d actually prefer a hand-painted poster.

Creativity doesn’t thrive on results alone. It thrives on personality. AI can replicate existing styles and combine them in seconds. But when everyone relies on the same models and the same aesthetic patterns, monotony sets in. The absurd thing is that we even sell this impoverishment as progress. We’re cutting out the graphic designer, the photographer, or the illustrator and getting more content than ever before in return. But more content doesn’t mean more culture. Perhaps we’re currently producing an endless amount of images, texts, and music that don’t move a single person.

AI can free us from monotonous work, accelerate research, and give people access to knowledge. It could indeed lead to greater prosperity and freedom. But that doesn’t happen automatically. If we leave its adoption solely to market forces, it will be used primarily where human labor can be eliminated most quickly. In that case, we won’t be creating a tool that works for us. We’ll be financing a system that privatizes the fruits of our labor and leaves the consequences of its rationalization to the general public.