Deutsche Bank's "Ten Thousand People Survey": Young people's anxiety about job impacts from AI far exceeds that of older colleagues

Author: Long Yue

Source: Hard AI

Artificial intelligence technology is reshaping the global labor market at an unprecedented speed, but the emotions it evokes among different groups of people are starkly different.

A research report released by Deutsche Bank on September 23 shows that young people, who are most familiar with the digital world, are the most anxious about the prospect of AI replacing their jobs; in contrast, older employees, who are perceived to be disconnected from new technologies, appear unusually calm.

This report, based on a survey of 10,000 employees in major economies in the U.S. and Europe, reveals profound generational, geographical, and trust gaps in the age of AI.

Are Young People's Jobs at Risk? The Generational Gap of AI Anxiety Has Emerged

The report data shows that the "employment anxiety" triggered by AI varies significantly across different age groups.

According to a survey conducted by dbDataInsights from June to August, as many as 24% of young employees aged 18 to 34 feel "very worried" about potentially losing their jobs to AI in the next two years (with a worry level score of 8-10, out of 10).

In contrast, only 10% of older employees aged 55 and above share the same level of concern.

This anxiety is not unfounded. The report cites recent studies from Stanford University and Harvard University to support this trend:

• Research from Stanford University shows that among occupations affected by AI, such as software engineering and customer service, the employment rate of young graduates aged 22-25 has decreased by 6% compared to the peak at the end of 2022.

• An analysis by Harvard University of U.S. resume and hiring data found that in companies using AI, the employment rate for entry-level positions has sharply declined, while the employment rate for senior positions has increased.

• Data from the U.S. Census Bureau also shows that the unemployment rate for recent college graduates (4.8%) has exceeded the overall unemployment rate for all workers (4.0%).

These data collectively paint a grim picture: the impact of AI on the labor market begins at the bottom of the "pyramid."

Geographical Differences in AI Adoption Rates

In addition to age, geographical location has also become a key dividing line in attitudes toward and adoption rates of AI.

The report points out that Americans generally express greater concern about AI replacing jobs compared to Europeans. Within the next two years, 21% of American respondents indicated they are "very worried," while this figure for Europeans is 17%. The report suggests that this may reflect the faster adoption of AI technology and higher social awareness in the U.S. market In practical application, the gap is even more evident:

• Adoption rate in work scenarios: The United States leads with 56%, followed by the United Kingdom at 52%, while Germany (41%), France, and Italy lag significantly behind.

• Adoption rate in home scenarios: Spain ranks first at 68%, higher than the average level of 57% among major European economies and the 59% in the United States.

This disparity indicates that while AI experiments at the individual level are occurring everywhere, the integration, governance, and policy support at the enterprise level are progressing faster in the United States and some European countries (like the United Kingdom), which may translate into productivity gaps between countries in the future.

A Huge AI Training Gap

In the face of challenges posed by AI, employees generally recognize the urgency of skill reshaping, but the readiness of enterprises and society is severely lacking.

According to the report, there is a strong demand for AI training among employees. 54% of American employees and 52% of European employees wish to receive AI-related training at work. However, there is a significant gap on the supply side: to date, only about one-third of American employees and one-quarter of European employees report having received any form of AI work training.

In the absence of adequate corporate training, some employees have begun to "self-rescue." About one-third of Americans and over one-quarter of Europeans learn about AI by watching videos, and about one-quarter read related articles. Nevertheless, half of the respondents indicated that they had not taken any measures for self-education in the past 3 to 6 months.

Trust Deficit: The Last Barrier to Large-Scale AI Application

Despite the initial applications of AI technology in personal assistants (with a usage rate of 61% in the U.S. and 50% in Europe), the report clearly states that "trust" is a key barrier preventing deeper and higher-value penetration into various fields.

When asked about their trust in AI in specific areas, the data reveals a general skepticism:

• Very low trust in high-risk areas: 40% of respondents do not trust AI to manage their personal finances, and 37% do not trust AI for medical diagnosis.

• Core functions are also questioned: 29% of people doubt the fairness of AI decision-making, and the same 29% do not trust the accuracy of information provided by AI.

• Common applications are not spared: Even customer service, one of the most common applications, faces distrust from 27% of respondents.

The report believes that although the accuracy of models is improving, the issues of "hallucinations" and biases brought about by the probabilistic nature of large language models (LLM) continue to erode user trust. Before establishing a reliable trust mechanism, the commercialization path of AI in industries that require high reliability, such as finance and healthcare, will be full of challenges.

This article is from the WeChat public account "Hard AI," for more cutting-edge AI news, please click here