The impact of AI is becoming evident! Employment growth in multiple industries in the United States has turned negative, and the unemployment rate among young employees in the tech industry has surged, but Goldman Sachs does not believe in a "mass unemployment wave."

The explosion of artificial intelligence, especially generative AI, has reignited concerns about "AI taking jobs."

On July 25th, according to the Chasing Wind Trading Platform, Goldman Sachs released a research report indicating that although the current impact of AI on the overall job market is still relatively limited, signs of layoffs and hiring contractions have already emerged in some industries most sensitive to AI, such as marketing, customer service centers, graphic design, search engines, and software development. Employment growth in these areas has turned negative.

Additionally, the technology sector's share of overall employment has also fallen below long-term trend levels, and the unemployment rate for young tech talent has risen by nearly 3 percentage points since the beginning of the year, confirming numerous real cases of AI affecting the employment of recent tech graduates.

However, Goldman Sachs also stated that while the impact of AI has already manifested in some industries, it will not lead to mass unemployment in the next decade. The main reason is that AI-related technological innovations will create a large number of new jobs, offsetting some of the job losses caused by automation.

Which Jobs Are Most at Risk? High-Risk Occupations in the Eyes of AI

Goldman Sachs noted that the current impact of generative AI on the overall macroeconomy and labor market is still relatively small, mainly because the overall adoption rate among companies remains low.

Overall, the vast majority of companies have yet to incorporate AI into their daily workflows, with only 9.3% of companies reporting that they have actually applied generative AI in their production processes in the past two weeks.

Goldman Sachs' model estimates that at the current pace of AI adoption, approximately 6%-7% of jobs may be replaced by AI in the coming years, with the replacement rate fluctuating between 3% and 14% under different assumptions. However, this does not mean that these individuals will never find work again.

Goldman Sachs estimates that for every 1 percentage point increase in productivity, the short-term unemployment rate will rise by 0.3 percentage points, but this impact will disappear after two years. AI will cause some people to lose their jobs, but the vast majority will ultimately transition to new employment, just as in past technological revolutions.

Additionally, Goldman Sachs referenced comments from corporate executives and academic research to identify the following five key risk factors that assess which occupations are more likely to be replaced by AI.

First, AI tools cannot be completely error-free, so for positions where errors could lead to significant financial or reputational losses, companies are more inclined to retain human workers to avoid risks.

Second, for highly repetitive and structured tasks, AI's automation capabilities are most pronounced, making these positions more susceptible to replacement.

Third, the closer the tasks are related, the more likely they are to be automated as a whole by AI.

Fourth, if the tasks exposed to AI in a certain occupation have low value and employees do not have a comparative advantage in these tasks, then while these tasks may be easily replaced, the overall position may be harder to completely replace.

Fifth, companies generally believe that the automation potential of backend positions is greater.

The results show that the jobs most likely to be replaced share the following characteristics: the work content is highly repetitive (such as data entry, meter reading, proofreading), belongs to logistical support roles (such as administrative assistants, customer service), has low costs associated with errors (such as copywriting, telemarketing), and the wage levels primarily come from tasks that AI can cover

In contrast, jobs that involve high communication, physical and cognitive integration skills, such as doctors, kindergarten teachers, and door-to-door salespeople, remain relatively safe.

AI is not a terminator, but a structural reshuffle

Goldman Sachs points out that the view that "technology will massively replace human labor" has been around for a long time, but historically, its accuracy has been very low. Despite continuous predictions that technological advancements will lead to persistent, structural unemployment, the total number of jobs has continued to grow in reality.

The main reason is that technology destroys old jobs on one hand, but simultaneously creates a large number of new jobs. These new jobs either directly arise from new industries spawned by technological innovation or come from the overall demand expansion driven by increased productivity due to technological advancements.

Goldman Sachs emphasizes that AI will not lead to "human extinction of employment"; it merely leads the labor market into a phase of structural reshaping. Looking back in history, 60% of jobs in the U.S. today did not exist in 1940. This means that while technology will destroy old jobs, it will also create new ones.

However, Goldman Sachs also warns that if the economy falls into recession during the transition period of AI popularization, total demand will decline, and companies may accelerate the replacement of human labor with AI to cut costs. This could lead to a significant spike in unemployment rates in the short term, with longer re-employment cycles as a large number of similar workers seek new jobs simultaneously, intensifying competition.

Conversely, if AI can create new profit growth points, it may also alleviate this shock, leading to a faster re-employment pace, thereby offsetting the impact of rising frictional unemployment