AI agents have already wiped out thousands of jobs over the past year, and the numbers don’t lie
Artificial intelligence and the job market rarely show up in the same sentence without making people uncomfortable.
And now there’s plenty of reason for that.
A new study from Goldman Sachs, led by economist Elsie Peng, dropped numbers that got a lot of people thinking — and rightfully so.
The takeaway is straightforward: automation powered by AI is already eliminating jobs at an accelerating pace, month after month, with no signs of slowing down.
According to the findings, over the past 12 months alone, AI-driven displacement has reduced job creation by 25,000 positions per month and pushed the unemployment rate up by 0.16%.
On the flip side, the use of AI tools by human workers — what experts call augmentation — generated around 9,000 jobs and helped bring unemployment down by 0.06%.
But the net result is still negative: 16,000 fewer jobs per month and a 0.1% upward pressure on unemployment.
The numbers are already on the table.
And the impact, as you’re about to see, isn’t hitting everyone equally. 👇
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Who’s feeling this impact the most?
The Goldman Sachs study makes it clear that the wave of automation isn’t hitting everyone with the same force. According to Elsie Peng, the negative effects fall disproportionately on less experienced workers. Early-career professionals who hold more operational, standardized roles are the ones on the front lines of this shift. And we’re not talking about some distant future — this is already happening right now, inside companies of all sizes, across sectors ranging from finance and accounting to customer service and logistics.
Artificial intelligence has evolved to the point where it can handle repetitive tasks based on data processing, information sorting, and administrative routines with a consistency and speed that simply has no comparison to the human pace. And that changes the math for companies when it comes to hiring.
Since ChatGPT debuted in 2022, sectors and occupations with high AI displacement rates have seen sharper declines in employment and more significant increases in unemployment, as Peng noted in the report. This data point matters because it shows the phenomenon isn’t a one-off — it’s a trend that has been building for over two years, gaining momentum with each new generation of language models and automation tools.
A recent employer survey conducted by Morgan Stanley reinforces this picture. Companies across five sectors considered most likely to face significant short-term impacts from AI adoption reported a net reduction of 4% in headcount. And here’s the most concerning finding: the number of roles eliminated and not backfilled was highest among entry-level professionals — those without prior experience.
On top of that, the geographic impact isn’t uniform either. Regions with economies more dependent on processing-based services — like secondary financial hubs, mid-sized cities with a strong corporate back-office presence, and areas with a concentration of outsourced jobs — tend to feel the pressure more intensely. Meanwhile, major tech centers continue absorbing specialized AI talent, creating a contrast that deepens already existing regional inequalities. The unemployment map is being redrawn in real time, and not randomly.
Mass layoffs are already a reality at big tech companies
If the macroeconomic data from Goldman Sachs paints a concerning picture, the concrete moves happening inside major tech companies turn that concern into something very tangible.
Companies like Block, Amazon, Oracle, and Meta have already carried out significant layoffs tied to AI adoption this year. The Block case drew particular attention: the company cut no less than 40% of its workforce, in a move directly linked to integrating AI agents into its operations.
Block’s CFO, Amrita Ahuja, addressed the decision publicly and offered a perspective worth reflecting on. According to her, the turning point happens when you realize you’ve automated a chunk of your work — something that used to take days can now be done in hours or less. Ahuja encouraged other executives and professionals to explore AI tools with curiosity, because hands-on experience is what creates that breakthrough moment of understanding the technology’s real potential.
That statement is telling because it represents the tone many corporate leaders are adopting: AI is no longer a matter of experimentation — it’s about execution. And when a company’s leadership takes that stance, the impact on headcount tends to be swift and deep.
Jeremy Allaire, CEO of Circle, was even more blunt in a recent interview during a New York Economic Club event. He stated that AI agents will replace a massive portion of work currently done by humans, at enormous scale. And according to Allaire, the most dramatic impact will hit office work — so-called white-collar jobs. But he also offered a counterpoint: professionals who embrace the capabilities of AI agents effectively gain new superpowers, and their ability to create impact grows dramatically.
This duality — the destruction of traditional roles and the creation of new capabilities for those who adapt — is what makes the current moment so complex and, at the same time, so full of possibilities.
The other side of the coin: augmentation and new opportunities
Not everything in the report points downward. Augmentation — the term used to describe using artificial intelligence tools as support for human work rather than a replacement — shows up as a real counterbalance within the data gathered by Goldman Sachs. The roughly 9,000 jobs generated monthly through this model show that when AI is used to expand human capacity instead of replacing it, the outcome can be positive for both productivity and the creation of new roles.
Professionals who learn to use these tools strategically can deliver more, with higher quality, in less time — and that has real market value.
Sectors like marketing, product development, design, research, and even medicine are seeing a new type of professional emerge: someone who knows how to work alongside AI, using it as an extension of their own skills. These professionals aren’t being replaced — they’re becoming more valuable precisely because they can get the most out of available tools without losing the critical judgment, creativity, and contextual intelligence that machines still can’t faithfully replicate. And that’s a real window of opportunity, especially for anyone already paying attention to market dynamics.
But it’s important not to romanticize this picture. The augmentation process requires access to training, time to adapt, technological infrastructure, and often institutional support that not every worker has. The transition isn’t automatic, and the rate at which new roles are created still can’t keep up with the speed at which others are being eliminated. The net loss of 16,000 jobs per month is the mathematical proof of that. So even though the augmentation path is promising, it needs to come with structured reskilling policies at scale — something that, so far, hasn’t arrived with the urgency it demands.
What the Goldman Sachs data reveals about the future
Economist Elsie Peng‘s study isn’t the first to map the relationship between automation and unemployment, but it’s one of the most detailed and up-to-date in terms of data granularity. The methodology used by Goldman Sachs is able to separate the impact of direct AI displacement from the impact of augmentation, which makes it possible to understand with more precision exactly where the market is being affected — and to what degree. That level of analysis is rare and represents an important step so that both companies and governments can make decisions based on concrete evidence, not just theoretical projections.
What the data reveals, in practice, is that artificial intelligence has already left experimental mode. It’s no longer being tested in corporate pilots or debated at tech conferences as a future trend — it’s operating now, within real workflows, making decisions, processing volumes of information impossible for human teams, and being integrated into critical business systems around the world.
And the job market is reacting to this in real time, with all the discomfort that kind of speed brings. The 0.16% pressure on unemployment might look like a small number in isolation, but when you multiply that effect over 12 months and project it across different economies, the cumulative volume starts to weigh in much more visibly.
What this report also implies — and this deserves close attention — is that the pace of AI adoption is likely to accelerate, not slow down. Increasingly capable language models, multimodal tools, autonomous agents, and AI-based decision-making systems are becoming more accessible and more powerful with every update cycle. That means the pressure on the job market described in the current data is, in all likelihood, just the beginning of a curve that hasn’t yet shown its peak. Understanding this dynamic today is essential for anyone who wants to navigate the job market with more confidence over the coming years. 📊
Between displacement and transformation: the market needs a response
The debate about AI and employment tends to fall into two extremes: either everything will be fine because new technologies have always created more jobs than they destroyed, or we’re heading toward a total collapse of the job market. The reality, as almost always, sits somewhere in the middle — but a middle that needs to be managed carefully and quickly.
The Goldman Sachs report doesn’t present doomsday scenarios, but it doesn’t downplay the problem either. It quantifies, and that’s exactly what turns a philosophical debate into an urgent public policy issue.
The historical logic that new technologies always generate new categories of work still holds, but with an important caveat: the gap between the destruction of old jobs and the creation of new ones tends to be painful for those living through the in-between. And with automation powered by artificial intelligence operating at a speed far greater than previous industrial revolutions, that gap could be longer and harder than other technological transitions we’ve already witnessed.
Professional reskilling programs, changes to the educational structure, incentives for companies that adopt augmentation models instead of full replacement, and more robust social safety nets are all key pieces of this puzzle.
The fact that a bank the size of Goldman Sachs is producing this type of analysis with such precision already says a lot about where the global financial market is focusing its attention. When institutions that move trillions of dollars in investment decisions start mapping AI’s impact on unemployment at this level of detail, it’s because the effects are already measurable and significant enough to influence long-term economic strategies.
The signal has been sent. The question now is what each of us — workers, companies, and governments — will do with this information. 🤖
