According to a survey by the Federal Reserve Bank of Atlanta, about 90% of corporate executives believe AI has not yet improved their company's productivity. More interestingly, factors that have driven corporate productivity increases since 2021 are not AI, but rather the efficiency gains brought by remote work during the pandemic. The two major actions taken by corporate management to enhance company value are layoffs and investing in AI, with the underlying logic being "AI makes employees more efficient, so fewer people can complete the same tasks." However, the data does not support this assumption.

The study cross-analyzed data from U.S. publicly listed companies over the past five years, covering millions of employee reviews, thousands of financial reports, and hundreds of AI investment and layoff announcements. The results showed that as the frequency of AI investment announcements increased, layoff announcements related to AI also rose in tandem. However, the market did not respond positively—stock returns averaged near zero after layoff announcements, and more than half of the related events had a negative market reaction.

Employees vote with their feet, and companies face the 'boomerang' effect

Analysis of millions of employee reviews on Glassdoor revealed that comments related to AI were significantly more negative than overall reviews. Four major concerns were concentrated on the risk of layoffs, insufficient training, limited opportunities for skill upgrades, and poor AI management by companies. After companies announced layoffs due to AI, employees' sentiments toward AI further declined, with the most intense anxiety about job security. In sharp contrast, management remained consistently optimistic about AI in about 10,000 earnings call transcripts. However, the study confirmed that this optimism was not significantly related to productivity. What truly determines whether AI can bring efficiency improvements is whether employees are willing to integrate it into their daily work—not just the slogans of executives.

Data from layoffs.fyi shows that over 122,000 people have been laid off in the tech industry in 2025, and another 126,000 in 2026. However, some companies are beginning to face the "AI boomerang"—rehiring employees who were previously laid off due to AI, because after implementing AI, they discovered that certain scenarios still rely on experienced "human knowledge banks." However, trust has already been broken, and rehired employees remain anxious about the next round of layoffs. This rift cannot be easily fixed with money. Additionally, the cost of hiring new people is also high; it often takes one to two times the salary of the former employee to train a new person properly. For senior positions, the hiring process can take three to six months. Worse still, many companies have destroyed their junior talent pipeline, leading to a lack of successors for mid-level positions. The cost of AI-related layoffs is far greater than what management initially calculated.