Goldman Sachs Raises AI Job Displacement Forecast to 15 Million Workers
Artificial intelligence is no longer a distant disruption on the horizon — it is actively reshaping the U.S. labor market, and Wall Street is taking notice. Goldman Sachs has significantly revised upward its estimate of the share of American jobs that could be displaced by generative AI, lifting its forecast from a previous range of 6–7% to over 9%. That figure translates to roughly 15 million U.S. workers potentially losing their jobs to AI-driven automation over the next decade, according to a new report from Goldman Sachs Economist Joseph Briggs.
The updated forecast has sent ripples through both economic circles and corporate boardrooms, raising urgent questions about workforce planning, policy responses, and the long-term human cost of the AI productivity boom. Yet the picture Goldman Sachs paints is not entirely bleak — the same report forecasts that AI will also be a powerful engine of job creation in the years ahead.
Why Goldman Sachs Revised Its Estimate Upward
The jump from a 6–7% displacement estimate to over 9% is not simply a matter of pessimism — it reflects a fundamental change in how Goldman Sachs is measuring AI's impact on employment. The previous estimate examined the existing "stock" of unemployed workers, essentially looking at a snapshot of joblessness at any given moment. The new methodology takes a more dynamic approach, analyzing the "flow" of workers out of existing jobs over time.
This flow-based model tracks how productivity gains driven by technology translate into real-world job destruction. Specifically, the current methodology assumes that each 1% increase in technology-driven productivity will cause a 0.5% to 0.6% increase in the job destruction rate over the following two years. Because generative AI is expected to deliver productivity gains that outpace earlier waves of automation, the updated model produces a meaningfully higher displacement estimate.
In practical terms, this means that as businesses adopt AI tools and systems that allow them to accomplish more with fewer human workers, those productivity gains will gradually but consistently push some workers out of their existing roles. The effect compounds over time, which is why Goldman Sachs frames the displacement as a 10-year trajectory rather than an immediate shock.
What Does 15 Million Displaced Workers Actually Mean?
Fifteen million is a number large enough to demand serious attention, but context matters enormously when interpreting what it means for the overall health of the U.S. labor market. Briggs notes that if job losses are distributed across a full decade — and if most displaced workers are able to find new employment within approximately one year — then the peak unemployment rate impact of AI would actually remain below 1%. That is a relatively modest macro-level disruption, even if the human stories behind that statistic are often painful and complex.
To put the scale in further perspective, the U.S. economy already generates between 25 million and 35 million new jobs every single year through natural churn, sector growth, and entrepreneurial activity. Goldman Sachs expects AI to add to that job creation figure over the long run, not simply subtract from it. The net effect, in the firm's view, is a labor market that is significantly transformed but not devastated — provided the transition is managed thoughtfully and workers have access to retraining and support.
The Jobs AI Is Creating: A New Frontier of Roles
While much of the public conversation around AI focuses on which jobs will disappear, an equally important story is emerging about the entirely new categories of work that AI is bringing into existence. Many of these roles did not exist even five years ago, and they are already generating significant demand across the technology sector and beyond.
Box CEO Aaron Levie recently highlighted one particularly striking example, writing on X that the role of forward deployed engineers — professionals who work directly at client sites to implement, customize, and optimize AI systems — is poised to become one of the most in-demand positions in tech. This role sits at the intersection of technical expertise and client-facing problem-solving, requiring skills that are difficult to automate and deeply human in nature.
Beyond forward deployed engineers, the AI economy is generating growing demand for roles such as:
- AI prompt engineers, who craft and refine the inputs that guide large language models toward useful and accurate outputs.
- AI ethics and compliance officers, who help organizations navigate the regulatory and reputational risks of deploying AI responsibly.
- Machine learning operations (MLOps) specialists, who manage the infrastructure that keeps AI systems running reliably at scale.
- AI trainers and data annotators, who provide the human-labeled data that makes AI models more accurate and context-aware.
- Human-AI collaboration designers, who architect workflows that allow human workers and AI systems to complement each other effectively.
These emerging roles share a common thread: they require human judgment, creativity, and interpersonal skills in ways that current AI systems cannot replicate. They also tend to be well-compensated, suggesting that the AI transition — while painful for some — may ultimately raise the quality and value of human work.
What Workers and Employers Should Do Now
The Goldman Sachs forecast is a call to action as much as it is a prediction. For individual workers, the most important takeaway is the urgency of building AI-adjacent skills, whether that means learning to work alongside AI tools in their current field or preparing for a transition into one of the many new roles the technology is creating. Waiting to see how things unfold is increasingly not a viable strategy.
For employers, the report underscores the importance of proactive workforce planning. Companies that invest in reskilling and upskilling programs now will be better positioned to retain institutional knowledge and avoid the costly disruption of rapid turnover. They will also be better prepared to fill the new AI-driven roles that their own technology adoption is helping to create.
For policymakers, the Goldman Sachs findings add urgency to conversations about social safety nets, education reform, and labor market support systems. A peak unemployment impact of under 1% is manageable — but only if the right infrastructure is in place to help displaced workers find their footing quickly.
The Bottom Line: Disruption With a Path Forward
Goldman Sachs' revised forecast of 15 million AI-displaced U.S. workers over the next decade is sobering, but it is not a verdict of doom. The same research that identifies the scale of displacement also points to robust job creation, modest macro-level unemployment impacts, and the emergence of entirely new categories of meaningful work. The AI transition will be uneven, difficult, and in many cases deeply personal — but it is also navigable. The key lies in acting now, investing in human adaptability, and ensuring that the extraordinary productivity gains AI promises are shared broadly across the workforce and society.