Companies Replaced Workers With AI, Then Rehired Them — At Brutally Lower Pay

Some companies that cut workers in favor of AI are now hiring people back. But the returning jobs don’t always look the same. In many cases, the pay is lower, the titles are different, and the terms are worse than before.

A survey found that 32% of U.S. hiring managers said a role eliminated because of AI was later reopened for the same or similar position. Another survey of 2,000 hiring managers put that number at 29%. The rehires often came after quality problems or customer complaints forced companies to reverse course.

Klarna, the payments company, became a well-known example. Its AI assistant reportedly handled 66% of customer conversations across 35 languages. The company reduced its workforce by about 30% through attrition and redeployment. But customer satisfaction declined. Klarna eventually brought human agents back for complex and empathy-heavy cases.

The pattern isn’t limited to one company. Around half of firms that replaced people with AI reportedly experienced some kind of rehiring reversal. Some did it quietly after service quality dropped. Others brought workers back as contractors or in redesigned roles with different salary terms.

In one case in China, a company allegedly offered a worker a demotion and a 40% pay cut, saying AI could do the job. The worker refused and was fired. A court later ruled the dismissal improper and awarded compensation.

The cost of these reversals can be high. Rehiring often costs more than keeping staff in the first place. Recruiting, onboarding, and training new workers add hidden expenses. The savings from layoffs can vanish once AI underperforms.

AI tends to handle routine, high-volume tasks well. But it struggles with nuanced, judgment-heavy work. Customer service, recruiting, and legal review are among the most common areas where automation was tried and then partially reversed. Roles that require empathy or negotiation are especially hard to fully replace.

Many companies have shifted to hybrid models. They keep AI for simple tasks and restore humans for exceptions. The broader trend suggests AI adoption often leads to partial substitution, not total replacement. And for workers who do return, the deal isn’t always fair. This dynamic is part of a wider concern, as 40% of employers expect to reduce staff in areas where AI can automate tasks, even as reversals show the limits of that strategy.