If AI takes over the job, what will be the value of humans?
A morning at 1 am, a senior project manager gazing at a computer screen, their phone rings. A generation of project plans generated by AI in just three minutes. After polishing the language and verifying the numbers, they transmit the file to stakeholders. As they settle back, a sense of desolation settles in. The weight of what their years of experience meant for the company has suddenly evaporated.
Their fear isn't just the prospect of joblessness - it's a more fundamental anxiety: if their accumulated skills can be so easily replicated, what proves their worth as human beings?
Throughout history, people have measured their value by what they can produce and what tasks they can perform. Assisting colleagues with Excel spreadsheets, calming customer complaints, and finding direction in vast data sets, they felt indispensable. But now, intelligence is being absorbed into the realms of machines and capital.
Humanity's value is under question. To answer this, they must consider labor value through three lenses: monetary worth, contribution to others and society, and a sense of personal dignity. Historically, these three factors moved in tandem. But in the AI era, they may diverge. Economic studies predict that as labor costs drop, companies will reevaluate labor costs.
Writing reports, reviewing legal documents, foreign language translation, and basic coding, once mastered through years of university and experience, can now be done by machines in mere seconds or minutes. When a person's time costs less than the cost of using an algorithm, the price of human labor changes. Economist Daron Acemoglu and Pascual Restrepo argue that automation is not just about increased productivity, but about the 'task allocation' between humans and machines.
Even if automation increases productivity, the proportion of work done by humans could decrease. Increased productivity does not automatically lead to higher wages for workers. Even if the time spent on writing business plans is halved, it does not mean an employee's salary will double. Instead, the remaining time can be used to demand more work or reduce the number of new hires.
The silent change occurs within jobs themselves. People continue to work as project managers and senior managers, but the responsibility and authority gradually devolve to systems. They are still employed but their expertise and authority diminish. Here are the first tremors of unease about labor value. Perhaps the real problem is not the number of jobs disappearing, but the declining value of skill.
Conversely, there are labor-intensive jobs that machines find difficult to mimic. Nursing aides for elderly residents, social workers in crisis situations, and daycare teachers who comfort crying children - these professionals have been supporting others' lives long before AI arrived. The unsettling paradox is that even though society needs these jobs, the market does not necessarily reward them highly.
If social consensus and policy do not support it, the claim that "humanity matters more in the AI era" may be nothing more than a sentimental expression. The real value of humanity may not rise automatically in the AI era. It requires social agreement and policy support. A more realistic perspective focuses on collaboration between humans and AI.
Rather than AI completely replacing jobs, it will change how tasks are handled. In 2025, the International Labour Organization estimates that one in four global workers will be in jobs that generate data for AI. The key, however, is not just substitution but transformation. Rather than AI completely replacing jobs, it will change the way work is done.
As of 2025, about one out of every four global workers will be involved in jobs that generate data for AI. The crucial point is not just substitution but transformation. Many jobs consist of various tasks, and some tasks require human intervention even after AI takes over.
Written by urgent.news from Hankyoreh's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.