Goldman Sachs: AI eliminates net 16,000 US jobs per month - yet 80 percent of companies see no productivity gains
What it really says
A CNN Business analysis published on May 10, 2026 synthesizes two seemingly contradictory research findings: On one hand, Goldman Sachs calculated from real payroll data that AI is destroying approximately 16,000 net jobs per month in the United States. In detail: roughly 25,000 positions are eliminated monthly through AI substitution, while about 9,000 new positions are created through AI augmentation. Over the past year, this sums to 192,000 lost jobs. Generation Z is disproportionately affected: employment in AI-exposed occupations has dropped 16 percent for workers aged 22-25. The most affected are routine office jobs - data entry, customer service, legal support, and accounting. On the other hand, a comprehensive NBER study surveying over 6,000 executives found that more than 80 percent of companies report no measurable productivity gains despite deploying AI. 90 percent of surveyed managers state that AI has had zero impact on headcount at their organization over the past three years. The CNN analysis concludes: AI is not replacing entire occupations but rather individual tasks within occupations - the transformation is proceeding more slowly and in a more fragmented manner than headlines suggest.
Our assessment
This story illustrates a paradox typical of the AI debate: on one side, real, measurable job losses; on the other, surprising inertia across the broader economy. Both data points are simultaneously true - they simply apply to different segments. The Goldman Sachs figures show that fear of AI-driven job loss is not unfounded, especially for career starters and young people in administrative roles. 16,000 net lost jobs per month is not a catastrophe scenario, but neither is it zero. At the same time, the NBER study shows that the revolution has not yet arrived in the vast majority of companies: most firms use AI (69 percent), but see no meaningful changes in either productivity or employment. This suggests we are in a transitional phase - the technology exists, but organizational restructuring takes time. For those affected, this is a double-edged sword: those losing their jobs today face a real problem; those worried about the future have more time to adapt than feared.
Relevance for Germany
The results must be transferred to Germany with caution, as the US labor market differs structurally from Germany's: stronger employment protection, worker co-determination, and the dual vocational training system all act as buffers. Nevertheless, the trends are relevant. Germany's Federal Employment Agency reports that job listings for classic office work declined 12 percent in 2025/2026. The ifo Institute estimates that up to 3.4 million workers in Germany hold positions that are highly AI-exposed. The Generation Z problem is visible in Germany as well: youth unemployment in academic professions has risen to its highest level since 2015 in 2026. At the same time, a Bitkom survey from March 2026 shows that only 34 percent of German companies productively deploy AI - confirming the NBER finding that transformation is proceeding more slowly than the debate suggests.
Fact check
The Goldman Sachs figures (16,000 net lost jobs/month, 192,000 over the past year) come from an analysis published in April 2026 based on US payroll data, consistently reported by Fortune, CNN, and other outlets. The NBER findings (80%+ with no productivity gains, 90% with no employment effect) are based on Working Paper 34984 by Barrero, Bloom, Davis and others, surveying over 6,000 executives. The Generation Z figure (16% employment decline in AI-exposed roles) comes from the Goldman Sachs analysis. Limitation: Goldman Sachs' methodology for causally attributing job losses to AI is not publicly verifiable in detail; other factors such as economic cycles and outsourcing may also play a role.
Source
- • CNN Business 10.05.2026 (cnn.com/2026/05/10/tech/ai-taking-jobs)
- • Goldman Sachs Economics Research April 2026 (goldmansachs.com/insights/articles/the-jobs-ai-is-likely-to-boost-and-those-it-may-disrupt)
- • Fortune 06.04.2026 (fortune.com/2026/04/06/ai-tech-displacement-effect-gen-z-16000-jobs-per-month/)
- • NBER Working Paper 34984: Artificial Intelligence, Productivity, and the Workforce (nber.org/papers/w34984)
- • NBER Working Paper 34836: Firm Data on AI (nber.org/papers/w34836)