The acceleration of enterprise AI applications continues to increase the risk of unemployment among workers

Goldman Sachs economist Pierfrancesco Mei pointed out in a report released on Tuesday that the impact of AI on the job market has entered an explicit stage, and job losses in industries deeply affected by AI are clearly visible. However, the overall loss rate is still at a moderate level and has not yet triggered large-scale employment turbulence. This judgment echoes Goldman Sachs' previous quantitative research on the impact of AI on employment - its previous calculations showed that AI will achieve automated restructuring of 25% of global working hours, but only 6-7% of positions will be permanently replaced, and the replacement process will be gradual, mostly through natural attrition, recruitment freezes, and other methods, rather than centralized layoffs.

Despite the current mild impact, Goldman Sachs remains cautious about future employment trends. The report emphasizes that the job substitution effect brought by AI will further manifest in 2026, with the most direct impact being a slight increase in global unemployment rate. Based on Goldman Sachs' overall forecast for the global economy in 2026, the US economy is expected to experience a "high growth, low employment" split that year, with GDP growth expected to reach 2.8%, far higher than market consensus. However, the unemployment rate will remain stable at around 4.5%, forming a typical "jobless growth" pattern. AI driven productivity improvement is one of the core drivers of this split, and companies are using AI to achieve "less manpower, higher output", gradually weakening the traditional correlation between economic growth and employment growth.

What is even more alarming is that there is a clear upward risk space for the impact of AI on unemployment. Pierfrancesco Mei added that if the adoption rate of AI by enterprises further accelerates in the future, and the scope of AI technology substitution continues to expand beyond current expectations, the increase in unemployment rate will far exceed the benchmark scenario, and the risk of job market turbulence will significantly increase. According to Goldman Sachs' calculations, under the benchmark scenario, the global unemployment rate will only increase by 0.5-0.6 percentage points during the AI transformation period. However, if AI applications accelerate, the peak unemployment rate may rise to 1.2-2.4 percentage points, and the impact will spread from the current high-risk industries to more areas.

From the perspective of industry distribution, the impact of AI substitution is not global coverage, but rather precise targeting of standardized, repetitive, and low creativity work processes, showing a clear structural differentiation. Among them, the pan service industry has been most significantly impacted, with AI automation penetration rates in customer service and administrative fields reaching 45% -50%. Basic positions such as regular agents, administrative assistants, and data entry have become the primary groups being replaced; The basic links of the financial industry are closely following, and the demand for manpower in positions such as junior audit and basic accounting continues to shrink; The fields of basic content creation and repetitive manufacturing processes are also facing enormous pressure from AI substitution. In contrast, industries such as construction and manufacturing that rely on physical operations, as well as fields such as healthcare and high-end creativity that rely on human empathy and complex decision-making, can still obtain a certain employment buffer in the short term.