To what extent does the AI really replace jobs? Goldman Sachs looks under the hood and has 3 take -out dishes to defuse media threw

For the second quarter of 2025, the Hatzius team found “notable progress” in the adoption of AI, with 9.2% of American companies using AI to produce goods or services, against 7.4% in the first quarter. The report also provides a nuanced image, noting that, although generative AI and related technologies quickly reshape the investment and productivity of businesses, their effect on employment evolves at a slower and more subtly pace.

Here are three dishes to remember from the Goldman AI Adoption Tracker.

1. Limited Labor Market Disturbance (so far)

Despite an increase in the adoption of AI in American companies, the research note has found that the overall labor market results remain largely not affected. In simple terms, “the impact of AI on the job market remains limited and there is no sign of a significant impact on most of the labor market results.” This contrasts with the signs that the technological sector reduces the jobs exposed to AI, and with several prominent CEOs warning AI could move more than 50% of the labor of white collars.

More specifically, Goldman claims that key measures such as employment growth, wage gains, unemployment rates and dismissal rates in the industries exposed to AI have shown little statistically significant difference compared to the less exposed sectors. AI job offers now represent 24% of all computer openings, but still represent only 1.5% of the total job offers, which indicates that if the technological roles adapt, the broader change in labor is progressive.

In particular, the unemployment rate of the professions exposed to AI has now been reconciled with the wider economy, refuting the early fears of mass movement. There was no recent announcement by explicitly invoking AI as a cause, further emphasizing the current confinement of disturbance towards specific functions rather than in whole industries.

Goldman Sachs

On the other hand, analysts noted that payroll growth continues to underperform in professions where AI has an anecdotal impact, as with the notable example of telephone call centers. It suggests that something is happening that is only whispering. However, it is early.

2. Productivity gains concentrated, but significant

Goldman says that the influence of AI on productivity in which it has already been deployed is pronounced. The Hatzius team has cited academic studies and corporate anecdotes indicating that the adoption products of generating AI, on average, an increase of 23% to 29% of labor productivity. Estimates vary, academic studies generating a median by 16% and on average of 23%, while company anecdotes produce a median of 30% and on average of 29%. However, this suggests tangible improvements for early adopters.

The sectors taking advantage of the most actively generative AI – information, finance and professional services – see the highest increases in productivity as companies go from experimentation to the integration of AI in their work Noyaus.

Goldman Sachs

Business managers and economists expect adoption to deepen and that more organizations accumulate AI in their infrastructure, the impact of global productivity will become more visible in macroeconomic data.

3. The history of employment of AI: always in its first chapters

A recurring theme in the Goldman Sachs analysis is that the effect of the use of AI is still developing. Although AI -related openings are developing, especially in IT, there is also an increase in demand for roles such as Automatic Learning Engineers and AI researchers. The surveys reflect that a substantial part of companies plans to hire for these skills.

Productivity improvements can possibly expand more industries, and “the intensity of the AI” (share of roles using an AI) remains the highest in the information and professional service technology sectors, signaling where future job changes could first materialize.

The report indicates that the current IA impact on the labor market is limited, but transformation seeds are sown. The increases in the adoption of companies’ AI, especially among large and medium -sized enterprises, indicate future productivity and role changes. But for the moment, the fears of a generalized loss of jobs induced by the AI seem overestimated – at least until a wider and deeper integration of technology with business processes occurs.

While companies continue to develop AI and support infrastructure mature, opportunities and challenges will be amplified, guaranteeing close observation by decision -makers, business leaders and workers.

Goldman Sachs refused to comment more.

For this story, Fortune Used a generative AI to help an initial project. An editor checked the accuracy of the information before the publication.