One of the biggest questions raised by artificial intelligence is how much jobs it will eliminate and when that job loss will begin. Anthropic CEO Dario Amodei famously predicted that AI could eliminate half of entry-level white-collar jobs by 2030. Others are more optimistic that, like past innovations, AI will eventually create as many jobs as it replaces.
But for many tech workers, the transformation has already begun. Salesforce cut 4,000 customer support employees last September, citing efficiency gains from agent AI. product. Goldman Sachs and Hewlett-Packard have made similar moves, with HP saying its AI efforts will result in up to 6,000 job cuts by 2028. Meanwhile, Duolingo has announced that it will no longer use human contractors for tasks that can be handled by AI, and will only hire for roles that cannot be further automated. And Klarna, which cut 40% of its workforce through an AI-powered hiring freeze, expects to cut a further 33% of its workforce by 2030.
What is AI cleaning?
Broadly speaking, AI washing refers to companies exaggerating or misrepresenting how extensively they use artificial intelligence. In the context of layoffs, this means blaming economically motivated layoffs on AI efficiencies that do not yet exist or have not been fully implemented.
Overall, employers announced more than 1.2 million layoffs in 2025, the most since 2020, and AI was cited in nearly 55,000 of those cuts, or 4.5%, according to research firm Challenger, Gray & Christmas. One of the biggest job cuts this year came from Amazon, which cut about 30,000 jobs across the company. CEO Andy Jassy initially touted the company’s use of generative AI and AI agents. As a means of reducing personnel, However, it later clarified that these reductions were “not actually AI-driven, at least not currently.”
Looking through quarterly earnings reports and investor calls, one might think that workers are already being replaced by AI tools that do the work for them. But in reality, AI’s current capabilities are often insufficient. of Executives at disruptive technology companies typically say this when justifying mass layoffs. Today’s systems can transcribe meeting notes, generate code, and automate back-end business processes, but they’re far from replacing entire departments, leading some skeptics to wonder if companies are just using AI as a convenient cover. Critics call this “AI washing.”
The rise of AI cleaning
Similar to misleading marketing techniques such as greenwashing and rainbow washing, “AI washing” was originally used to criticize companies that claimed to use artificial intelligence when in fact they did not. But more recently, the term has come to refer to companies disingenuously blaming AI to explain layoffs and other unpopular business decisions, when the overall picture is much more complex.
In its latest AI employment impact forecast, market research firm Forrester said, “Layoffs due to economic reasons are being confused with layoffs due to AI.” “Many companies announcing AI-related layoffs do not have mature, vetted AI applications ready to fill those roles. This highlights a trend of ‘AI cleaning,’ which results in financially motivated layoffs against future AI deployments.”
Other experts suggest that AI-driven job cuts may be premature and misleading. In a separate survey, 60% of more than 1,000 executives surveyed said they had cut staff. Looking forward to it Improve AI efficiency. A further 29% said they were hiring fewer people than usual. Only 2% said they had made significant headcount reductions as a result of implementing AI.
This practice has been around for years, but has gained renewed attention in recent months. The term “AI washing” was also used by OpenAI CEO Sam Altman., They agreed that some companies blame technology for unrelated layoffs.
Additionally, the AI cleaning trend is likely to continue until 2026. Nearly 60% of U.S. recruiters surveyed by Resume.org said they plan to implement layoffs in 2026, with AI or automation being the most cited reason. At the same time, only 9% said AI has completely replaced certain roles, and 45% said it has partially reduced the need for new hires. Nearly 60% said they value AI’s role in job cuts and layoffs because AI is viewed more favorably than financial constraints.
Why do companies use AI cleaning?
Peter Cohan, an associate professor of management practice at Babson College, told Built-In that employees cite AI in layoff announcements because it’s “the worst reason a company can use.”
For example, if companies blame tariffs on layoffs, they risk being attacked by the Trump administration. If a decline in sales is cited as the reason, investors may sell their shares. But by blaming AI, the company can appear more advanced and innovative, essentially turning bad news into good news. Cohan noted that hot topics such as the pandemic, tariffs and AI are often used in earnings calls to explain business challenges, and said they can be “safe things to hide behind.”
Some experts have suggested that AI job cuts could be caused by right-sizing the high-tech sector’s labor market. During the post-pandemic jobs boom of 2021, technology companies competed with each other to attract venture capital and growth-minded, career-hopping tech professionals. The ensuing market contraction forced expansion plans to end, leaving many companies with overemployment.
“Many organizations over-hired during and after the pandemic. There was never enough demand to justify it,” Manish Jain, principal research director at Infotech Research Group, told Built In. “Many, if not all, are now willing to use AI as a surface to lighten the burden for large numbers of people.”
Is AI actually to blame here?
Artificial intelligence appears to be slowing hiring in certain sectors of the labor market, particularly customer support, software engineering, and entry-level white-collar jobs. A Stanford University study found that employment of early-career workers in AI-exposed occupations has declined by 16 percent since the introduction of ChatGPT in 2022, suggesting that, at least for the time being, the technology is reshaping the bottom rungs of the career ladder rather than hollowing out entire sectors.
But zoom out and the broader job market looks less turbulent. A study by the Yale Institute for Budget Research found no overall change in employment for workers in jobs exposed to AI. A National Bureau of Economic Research study similarly found that AI has so far had little measurable impact on employment or productivity. However, respondents said they expected productivity to rise by 1.4% and employment to fall by 0.7% over the next three years, which is still far from the shockwave some industry leaders have suggested. The Oxford Economics study also said that the 55,000 AI-related job cuts cited in the Challenger, Gray & Christmas report accounted for only 4.5% of the total layoffs, and that the weak job market for young professionals could be due to an “oversupply” of young professionals with university degrees.
It may be a bit premature for companies to cite AI as a reason for layoffs. Chatbots like ChatGPT generated a lot of hype, but it wasn’t until a few years ago that most companies started integrating chatbots at an organizational level. It has taken time for companies to understand the capabilities of AI and implement these technologies in a secure and reliable manner. Even today, most organizations are still wary of completely outsourcing business functions such as customer service and legal support to AI.
“Many of the implementation details have to be completely bulletproof before anyone commits their life to it,” Cohan said. “There is a huge gap between the ease of saying something and the difficulty of making it happen.”
However, this does not mean that AI lacks capabilities. Max Leaming, head of data science and AI solutions at Manpower Group, told Built In that “it has more capabilities than most people not only realize, but are willing to admit.” Thanks to Vibe Coding, software engineers can now build and test apps in about an hour, a task that would normally take three weeks. And AI tools are making inroads into areas such as financial analysis and legal research. More professionals are starting to adopt this technology to increase productivity, but productivity is difficult to quantify, Leeming says.
While AI may be effective in enhancing certain roles and driving efficiencies here and there, Cohan said it’s still “not clear,” at least for now, how AI will create what he calls a “quantum leap in value” that will prompt companies to reorganize on a large scale.
“Right now, these uses of AI seem like a cost-reducing thing, but the value generator is…the growth that exceeds expectations,” Cohan said. “With a few exceptions, we don’t really see companies leveraging AI to generate growth at scale that exceeds expectations,” he added, adding that Google and Meta are leveraging AI to drive revenue growth in their core advertising businesses.
Will AI cause further layoffs in the future?
None of this is to say that AI won’t displace, restructure, or completely eliminate millions of jobs in the future.
Forrester predicts that this technology will eliminate 10.4 million jobs by 2030. This is a 6% reduction. This estimate exceeds the job losses seen during the Great Recession of 2008, but is somewhat modest compared to other estimates. A 2023 report from Goldman Sachs estimates that 300 million jobs will be exposed to AI by 2030, but that most of those jobs will be replaced by automation, not necessarily replaced. Meanwhile, a World Economic Forum report predicts that AI will displace 92 million jobs by 2030, but also create 170 million jobs, resulting in a net addition of 78 million jobs, an increase of 7 percent from current levels.
With hundreds of millions of jobs at stake, some federal lawmakers are calling for more concrete data on how AI will impact the workforce. A bill proposed in the U.S. Senate aims to quantify the impact of AI on the workforce by requiring federal agencies, public companies, and some private companies to provide data on how many roles are eliminated, added, or left unfilled because of AI. Employers will also need to quantify the number of employees who will be retrained due to AI implementation. The bill, which has not yet been scheduled for a vote, is one of several efforts underway to better understand how AI will impact the labor market.
However, identifying the impact of AI on employment can be complicated for a variety of reasons. This is not only difficult to quantify, but also relies on self-reporting by companies, which have their own motivations. Since March 2025, New York State gave employers the option to cite “technological innovation or automation” in legally required layoff notices. wired Of the 160 companies that filed notices (including companies such as Amazon and Goldman Sachs that cited other AI efficiencies in communications), not a single company checked the box citing AI-driven workforce reductions.
While we may not have precise data on the real-world impact of AI, it is clear that this technology has not yet had the full impact of removing humans from the workforce completely. And for executives to “wash out” layoff announcements with AI may reveal that they see AI as a means to reduce jobs, which could lead to workers distrusting and potentially thwarting future AI implementation plans.
Is AI causing mass layoffs?
It’s not wide. While some companies have cited AI in their layoff announcements, AI-related layoffs will only account for about 4.5% of total layoffs in 2025, according to Challenger, Gray & Christmas. Most evidence suggests that the impact of AI on labor so far will be limited and uneven.
Why are companies blaming AI for layoffs instead of other factors?
AI can be a safer and more favorable explanation than alternatives such as weak demand, overemployment, or business failure. Attributing layoffs to AI could signal innovation to investors and avoid political or market backlash.
Is AI having a bigger impact on certain types of jobs than others?
Indeed, early-career and entry-level white-collar workers, especially those in customer support and software programming, are more likely to be exposed to early AI-related hiring and attrition slowdowns.
