Does AI Lower Wages?—A Survey of Over 3,000 German Companies Reflects a "Tougher Future" for Young Workers in the Labor Market

Since the advent of generative AI, the debate over whether AI will eliminate jobs has been repeated worldwide.

However, the changes occurring in the actual labor market may not be as simple as a binary choice between having a job or not.

If AI enables the same work to be done with fewer people, the wages companies are willing to pay each worker may also change. More importantly, the impact will not be equal for all workers.

A survey conducted by Germany's ifo Institute in June 2026 clearly illustrates this possibility.

The survey targeted over 3,000 German companies already using AI, asking how they believe AI usage will affect employee wages over the next five years.

The results do not suggest a future where AI uniformly lowers wages for all workers.

Rather, there is a possibility of widening wage disparities between younger and more experienced workers, and between low-skilled and high-skilled workers.

About Half of Companies Expect "Wages for Young Workers to Fall"

The most noteworthy aspect is the companies' views on workers with short work histories.

About half of the companies believe that the wages of workers with less than five years of experience will decrease due to AI usage.

In contrast, about 40% of companies predicted wage declines for workers with more than five years of experience.

In other words, companies view the impact of AI not as something affecting "all human jobs" but as varying based on workers' experience and skills.

This difference is likely not coincidental.

The tasks initially assigned to young employees in companies often include document preparation, information gathering, data organization, drafting, simple analysis, and presentation creation.

These are also areas where generative AI is rapidly excelling.

Information that a new employee would have taken hours to research can now be organized by AI in minutes.

Tasks such as creating drafts for sales materials, writing emails, summarizing meeting content, and organizing market information can now be automated or semi-automated.

As a result, the economic value of tasks once entrusted to "young employees" may decrease from a company's perspective.

Even if AI does not completely replace new employees, a reduction in the labor hours required per new hire could exert downward pressure on hiring numbers and wages.

The Impact on Young Workers is Particularly Significant in the Service Industry

When viewed by industry, even more intriguing differences emerge.

For workers without a university or specialized college degree, the percentage of companies in the service industry that believe wages for those with less than five years of experience will decrease due to AI is 53.3%.

Even for those with more than five years of experience, it reaches 44.2%.

In the commercial sectors such as retail and wholesale, the figures are 47.9% and 41.3%, respectively.

In manufacturing, they are 46.5% and 38.9%.

Meanwhile, in the construction industry, the figures are 39% and 31.3%, indicating fewer companies expect wage declines compared to other industries.

This difference may reflect the distinction between jobs that are easily replaceable by AI and those that are not.

Jobs that are completed solely within the digital space are more easily automated by AI.

This is especially true for tasks involving text, data, documents, and images.

On the other hand, tasks requiring physical space work, on-site judgment, and human coordination, such as those on construction sites, are difficult to replace entirely with generative AI.

Of course, advances in robotics and automated construction technology could change the situation, but at least the current generative AI boom's impact varies significantly by industry.

It's Not Simply a Matter of "Being Safe with a High Degree"

Interestingly, even for workers with university degrees, a significant number of companies predict wage declines for young workers.

This means that the structure is not "immune to AI's impact if one has received higher education."

However, there is a significant difference.

For workers with degrees, the percentage of companies predicting wage increases due to AI is higher.

Particularly for highly qualified personnel with more than five years of experience, 26% of companies anticipate wage increases due to AI.

Even for those with less than five years of experience, the figure is 16.3%.

In contrast, only 5.9% of companies expect wage increases for workers without degrees and less than five years of experience.

These figures suggest that rather than automatically benefiting highly educated personnel due to AI, the value of personnel who can enhance productivity using AI, equipped with expertise and experience, may rise.

For example, if an experienced specialist can use AI to handle analyses and document preparation previously done by a team, that specialist's productivity significantly increases.

Companies may choose to reduce numbers while paying higher wages to key personnel.

As a result, the disparity among workers becomes more critical than how the "average wage" moves.

Is AI Creating "Wage Polarization" Rather Than "Unemployment"?

In discussions about AI and employment, the focus often falls on the number of jobs AI will take away.

However, this survey indicates another important possibility.

Even if the jobs themselves remain, the wage structure may change.

For example, consider a company with 100 white-collar workers.

Even after AI is introduced, not all 100 may be laid off.

However, if the market value of 20 people making complex decisions using AI rises while the rarity of work done by 80 people following AI instructions decreases, wage disparities will widen.

AI may become a technology that separates those who can enhance value using AI from those who compete with AI, rather than a "human vs. machine" structure.

This is the significant implication of this survey.

In the Construction Industry, There is Also Hope for Wage Increases for Experienced and Highly Qualified Personnel

The construction industry is particularly intriguing.

For personnel with degrees, 24% of construction companies predict wage increases due to AI for those with less than five years of experience, and 29.8% for those with more than five years.

For highly qualified personnel with more than five years of experience, this is particularly high among the major industries surveyed.

In the service industry, the figure is 27.3% under the same conditions, 24.9% in commerce, and 24.2% in manufacturing.

This result suggests that being in an industry less likely to have jobs taken by AI is not the only important factor.

Humans who understand complex real-world situations, possess expertise, and make final judgments using AI information may become more valuable than before.

As AI handles calculations and document preparation, humans can focus on more advanced aspects such as design, judgment, negotiation, responsibility, and customer interaction.

The Question Spreading on Social Media: "How Will Young Workers Gain Experience?"

A particularly noted issue on social media regarding this survey is a problem separate from wages themselves.

 

That is the "entry point of experience."

On LinkedIn, there is a question about how young workers will gain five years of experience if AI takes over tasks such as research, document preparation, writing, and data analysis, which they previously handled to gain experience.

This is a very important issue.

Even those called veterans in companies did not start with advanced judgment abilities.

They gained experience through years of starting with simple tasks, failing, being corrected by superiors, and receiving feedback from clients.

However, if companies prioritize efficiency and delegate all "beginner tasks" to AI, productivity may increase in the short term.

But this could lead to a shortage of personnel capable of making advanced judgments in ten years.

Social media also reflects the opinion that "producing outputs completed by AI is different from having real experience and judgment."

Even if AI can produce excellent outputs, it does not necessarily mean one can explain why the answer is correct or where the risks lie.

In the era of AI, this distinction will be extremely important for human resource development.

"It's Not That Jobs Disappear, But the Nature of Newcomers' Jobs Changes"

On the other hand, there is a more optimistic view on social media.

The opinion is that "it's not that the jobs for newcomers themselves will disappear, but rather the nature of existing newcomer tasks will change."

If newcomers previously conducted information gathering and document preparation by themselves, in the future, they may increasingly delegate tasks to AI, check, correct, and combine the results.

Thus, what is required is not just "execution" but "orchestration."

What should be entrusted to AI?

Which information should be trusted?

Where should humans verify?

How should multiple AI tools and human experts be combined?

These abilities themselves could become new professional experiences.

From this perspective, AI does not completely take away growth opportunities for young workers.

However, companies need to redesign their systems for training newcomers.

Simply instructing them to "finish work quickly using AI" will not easily develop their judgment abilities.

New training is needed, such as explaining why they chose a particular answer, identifying issues in AI's responses, and experiencing customer interactions.

Will "Soft Skills" Become the Hard Skills of the AI Era?

Another prominent discussion on social media is whether the value of abilities only humans can possess will increase.

A LinkedIn post suggests that as AI supports research, analysis, writing, presentation, and programming, abilities related to emotions and sociality will become more important than ever.

Listening to people.

Understanding different perspectives.

Mediating conflicts.

Making judgments in uncertain situations.

Persuading clients.

Taking responsibility.

Leading a team.

These abilities have long been called "soft skills."

However, in a world where AI handles massive knowledge processing, these may become the "hard skills" that determine human market value.

However, there is also a warning on social media that emphasizing only human abilities while neglecting expertise is dangerous.

To critically evaluate AI's responses, one must understand the field in the first place.

It is difficult for someone unfamiliar with medicine to accurately evaluate AI's diagnosis.

It is also not easy for someone unfamiliar with law to spot issues in AI-generated contracts.

Therefore, what is required in the AI era is not a choice between expertise or human skills.

Combining expertise, AI utilization skills, judgment, and interpersonal skills becomes important.

For Companies, It's Not Just About "Paying Young Workers Less"

From the company's perspective, AI may appear as an attractive technology for reducing labor costs.

If the same work can be processed in a shorter time, companies can reduce staff or curb hiring.

However, another problem arises in the long term.

Reducing new hires will also reduce future mid-level employees.

If all work for young workers is automated, opportunities to gain experience will also decrease.

As a result, in a few years, there may be a shortage of "experienced individuals capable of making advanced judgments while using AI."

In other words, companies must balance short-term cost reduction through AI with long-term human resource development.

AI introduction is not just an IT department issue; it is becoming a matter of human resources, recruitment, education, and evaluation systems.

For Workers, "Being Able to Use AI" Will No Longer Be a Differentiator

The situation is complex for individuals as well.

A few years ago, "being able to use generative AI" was a special skill.

However, as AI usage becomes commonplace across companies, that ability may become a basic skill, akin to using a computer or search engine.

Thus, what becomes important is not "whether you can use AI" but "what you can do with AI."

##HTML