How much is AI really replacing jobs? Goldman Sachs looks under the hood and has 3 takeaways to defuse the hype

For the second quarter of 2025, Hatzius’ team found “notable progress” in AI adoption, with 9.2% of U.S. companies using AI to produce goods or services, compared to 7.4% in the first quarter. The report also delivers a nuanced picture, finding that while generative AI and related technologies are rapidly reshaping corporate investment and productivity, their effect on employment is evolving at a slower, subtler pace.

Here are three takeaways from the Goldman AI Adoption Tracker.

1. Limited labor market disruption (so far)

Despite a surge in AI adoption across U.S. firms, the research note found overall labor market outcomes remain largely unaffected for now. Simply put, “AI’s impact on the labor market remains limited and there is no sign of a significant impact on most labor market outcomes.” This contrasts with signs that the tech sector is cutting jobs exposed to AI, and with several prominent CEOs warning AI could displace upward of 50% of the white-collar workforce.

Specifically, Goldman says key metrics such as job growth, wage gains, unemployment rates, and layoff rates in AI-exposed industries have shown little statistically significant deviation from less exposed sectors. AI-related job postings now account for 24% of all IT openings, but still represent just 1.5% of total job postings, indicating that while technology roles are adapting, the broader workforce shift is gradual.

Notably, the unemployment rate for AI-exposed occupations has now reconciled with the wider economy, refuting early fears of mass displacement. There have been no recent layoff announcements explicitly citing AI as the cause, further underscoring the current containment of disruption to specific functions rather than entire industries.

On the other hand, the analysts noted, payrolls growth continues to underperform in occupations where AI is having an anecdotal impact, as with the notable example of telephone call centers. This suggests that something is happening that is only being whispered about. Still, it’s early days.

2. Productivity gains concentrated, but significant

Goldman says AI’s influence on productivity where it’s already been deployed is pronounced. Hatzius’ team cited academic studies and company anecdotes indicating generative AI adoption delivers, on average, a 23%–29% boost to labor productivity. The estimates vary, with academic studies generating a median of 16% and average of 23%, while company anecdotes produce a median of 30% and average of 29%. Still, this suggests tangible efficiency improvements for early adopters.

Sectors leveraging generative AI most actively—information, finance, and professional services—are seeing the largest increases in productivity as firms move from experimentation to integrating AI into their core workflows.

Business leaders and economists expect that as adoption deepens and more organizations build AI into their infrastructure, the aggregate productivity impact will become more visible in macroeconomic data.

3. The AI employment story: still in its early chapters

A recurring theme in the Goldman Sachs analysis is that the full employment effect of AI is still developing. While AI-related openings are growing, especially in IT, there is also an uptick in demand for roles such as machine-learning engineers and AI researchers. Surveys reflect that a substantial share of companies are planning to hire for these skillsets.

Productivity improvements may eventually widen to more industries, and “AI intensity” (share of roles heavily using AI) remains highest in information-technology and professional-service sectors, signaling where future employment shifts might first materialize.

The report said the current impact of AI on the labor market is limited, but the seeds of transformation are being sown. Increases in corporate AI adoption, especially among large and medium-sized firms, point toward future productivity and role changes. But for now, fears of widespread AI-induced job loss appear overstated—at least until broader, deeper integration of the technology with business processes occurs.

As companies continue to scale AI and as supporting infrastructure matures, opportunities and challenges will both be amplified, warranting close observation by policymakers, business leaders, and workers alike.

Goldman Sachs declined to comment further.

For this story, Fortune used generative AI to help with an initial draft. An editor verified the accuracy of the information before publishing.