The topic of AI taking jobs has been discussed for years, but the numbers have always been estimates, with predictions from different institutions varying so wildly that some suspect fabrication.
Goldman Sachs recently did something valuable: it quantified the actual monthly number of jobs AI is eliminating into a concrete figure.
A net loss of 16,000 jobs per month.
How the Number Was Calculated
Goldman economist Elsie Peng combined U.S. labor market data with AI exposure indices, breaking down two effects: AI substitution and AI augmentation.
| Effect | Monthly Impact |
|---|---|
| AI Substitution (jobs lost) | 25,000 jobs lost per month |
| AI Augmentation (new jobs created) | 9,000 new jobs per month |
| Net Change | Net loss of 16,000 jobs per month |
The methodology combined a standard AI exposure score with a complementarity index developed by the IMF to distinguish between two types of roles:
- Substitutable roles: Jobs where AI can handle most core tasks (insurance claims, billing, data entry)
- Augmentable roles: Jobs requiring human judgment, physical presence, or specialized knowledge (law, engineering management, healthcare)
The figures are based on actual data from the past year, not predictions.
Which Jobs Are Disappearing
The most affected are repetitive white-collar jobs:
- Data entry
- Customer support
- Paralegal-level support work
- Billing and claims processing
- Administrative tasks
These roles share a common characteristic: fixed tasks, no need for professional judgment, and describable in text or spreadsheets. The AI replaceability rate is not theoretical; it is substitution already happening.
A counterintuitive conclusion: manual labor jobs are currently safer. Work requiring physical presence, such as construction management, medical operations, and on-site services, has no effective AI replacement path yet. White-collar workers are being hit before blue-collar workers, contrary to many people's intuition.
Why Gen Z Is Bearing the Brunt
Goldman's analysis provides a specific figure: for each standard deviation increase in the AI exposure index, the wage gap for young workers widens by 3.3 percentage points.
The logic chain is clear:
- Young people typically enter entry-level roles — data entry, customer service, junior analysis
- These roles have the highest AI substitution rates
- Young workers lack accumulated professional judgment as a buffer, and their experience is insufficient to quickly transition to roles requiring human expertise
- Job competition is more intense, and bargaining power is weaker
This aligns with the Stanford 2026 report mentioning a 20% reduction in programmer positions for 22-year-olds, forming a logical chain: it's not that senior employees are laid off first; the entry barrier for newcomers is directly blocked. Senior employees have an experience moat; new graduates do not.
Are the Numbers Underestimated?
Goldman itself acknowledged the study's limitations: the data does not fully account for new jobs created by AI infrastructure construction, such as data center building, power system expansion, and hardware manufacturing.
If these were included, the net elimination figure would be smaller.
But there is a key issue here: data centers create engineering jobs, while eliminating customer service jobs. The skill requirements, salary levels, and affected populations of these two groups have almost no overlap. A balance in employment data does not help the 16,000 people affected each month; they will not find new jobs just because a new server farm is built in Dallas.
How Companies Are Responding
Some companies' approaches are worth noting:
BetterUp (an AI augmentation case): When AI took over scheduling tasks, employees were reassigned to in-depth candidate feedback and role sourcing — preserving jobs and moving people to areas AI cannot replace.
Snap, Atlassian, Block: Used AI progress as a justification for layoffs. Criticism has emerged in HR circles — using AI as an excuse for layoffs without a supporting redeployment plan harms both employees and the organization.
A Fortune 500 CHRO told Fortune bluntly: "We didn't have enough strategic consideration when we conducted the layoffs."
The Conclusion Is Direct
A net loss of 16,000 per month, nearly 200,000 jobs per year. Not a huge absolute number, but it is moving consistently in one direction with no signs of slowing down.
What deserves more attention than the total is the distribution: the jobs disappearing are white-collar entry-level roles, while the jobs being added are engineering and specialized roles. The two groups barely overlap. This structural shift in the labor market is not something that can be solved simply through retraining.
Sources: CocoLoop, AI is cutting 16,000 U.S. jobs a month — and Gen Z is taking the brunt, Goldman Sachs says (Fortune); Top HR leaders warn against using AI as cover for mass layoffs (Fortune)